Merge pull request #2975 from annando/1611-frontend-worker

The worker can now run from the frontend as well

boot.php

@@ -1385,6 +1385,10 @@ class App {
 
 	function proc_run($args) {
 
+		if (!function_exists("proc_open")) {
+			return;
+		}
+
 		// Add the php path if it is a php call
 		if (count($args) && ($args[0] === 'php' OR !is_string($args[0]))) {
 

doc/htconfig.md

@@ -29,6 +29,8 @@ Example: To set the directory value please add this line to your .htconfig.php:
 * disable_email_validation (Boolean) - Disables the check if a mail address is in a valid format and can be resolved via DNS.
 * disable_url_validation (Boolean) - Disables the DNS lookup of an URL.
 * event_input_format - Default value is "ymd".
+* frontend_worker (Boolean) - Activates the frontend worker which acts as a replacement for running the poller via the command line.
+* frontend_worker_timeout - Value in minutes after we think that a frontend task was killed by the webserver. Default value is 10.
 * ignore_cache (Boolean) - For development only. Disables the item cache.
 * like_no_comment (Boolean) - Don't update the "commented" value of an item when it is liked.
 * local_block (Boolean) - Used in conjunction with "block_public".

include/poller.php

@@ -15,7 +15,7 @@ use \Friendica\Core\PConfig;
 
 require_once("boot.php");
 
-function poller_run(&$argv, &$argc){
+function poller_run($argv, $argc){
 	global $a, $db;
 
 	if(is_null($a)) {
@@ -35,8 +35,6 @@ function poller_run(&$argv, &$argc){
 
 	$a->start_process();
 
-	$mypid = getmypid();
-
 	if ($a->max_processes_reached())
 		return;
 
@@ -53,14 +51,7 @@ function poller_run(&$argv, &$argc){
 	}
 
 	if(($argc <= 1) OR ($argv[1] != "no_cron")) {
-		// Run the cron job that calls all other jobs
-		proc_run(PRIORITY_MEDIUM, "include/cron.php");
-
-		// Run the cronhooks job separately from cron for being able to use a different timing
-		proc_run(PRIORITY_MEDIUM, "include/cronhooks.php");
-
-		// Cleaning dead processes
-		poller_kill_stale_workers();
+		poller_run_cron();
 	} else
 		// Sleep four seconds before checking for running processes again to avoid having too many workers
 		sleep(4);
@@ -69,57 +60,85 @@ function poller_run(&$argv, &$argc){
 	if (poller_too_much_workers())
 		return;
 
-	$cooldown = Config::get("system", "worker_cooldown", 0);
-
 	$starttime = time();
 
 	while ($r = poller_worker_process()) {
 
-		// Quit when in maintenance
-		if (get_config('system', 'maintenance', true))
+		// Count active workers and compare them with a maximum value that depends on the load
+		if (poller_too_much_workers()) {
 			return;
+		}
+
+		if (!poller_execute($r[0])) {
+			return;
+		}
+
+		// Quit the poller once every hour
+		if (time() > ($starttime + 3600))
+			return;
+	}
+
+}
+
+/**
+ * @brief Execute a worker entry
+ *
+ * @param array $queue Workerqueue entry
+ *
+ * @return boolean "true" if further processing should be stopped
+ */
+function poller_execute($queue) {
+
+	$a = get_app();
+
+	$mypid = getmypid();
+
+	$cooldown = Config::get("system", "worker_cooldown", 0);
+
+	// Quit when in maintenance
+	if (get_config('system', 'maintenance', true)) {
+		return false;
+	}
 
 	// Constantly check the number of parallel database processes
-		if ($a->max_processes_reached())
-			return;
+	if ($a->max_processes_reached()) {
+		return false;
+	}
 
 	// Constantly check the number of available database connections to let the frontend be accessible at any time
-		if (poller_max_connections_reached())
-			return;
-
-		// Count active workers and compare them with a maximum value that depends on the load
-		if (poller_too_much_workers())
-			return;
+	if (poller_max_connections_reached()) {
+		return false;
+	}
 
 	$upd = q("UPDATE `workerqueue` SET `executed` = '%s', `pid` = %d WHERE `id` = %d AND `pid` = 0",
 		dbesc(datetime_convert()),
 		intval($mypid),
-			intval($r[0]["id"]));
+		intval($queue["id"]));
 
 	if (!$upd) {
-			logger("Couldn't update queue entry ".$r[0]["id"]." - skip this execution", LOGGER_DEBUG);
+		logger("Couldn't update queue entry ".$queue["id"]." - skip this execution", LOGGER_DEBUG);
 		q("COMMIT");
-			continue;
+		return true;
 	}
 
 	// Assure that there are no tasks executed twice
-		$id = q("SELECT `pid`, `executed` FROM `workerqueue` WHERE `id` = %d", intval($r[0]["id"]));
+	$id = q("SELECT `pid`, `executed` FROM `workerqueue` WHERE `id` = %d", intval($queue["id"]));
 	if (!$id) {
-			logger("Queue item ".$r[0]["id"]." vanished - skip this execution", LOGGER_DEBUG);
+		logger("Queue item ".$queue["id"]." vanished - skip this execution", LOGGER_DEBUG);
 		q("COMMIT");
-			continue;
+		return true;
 	} elseif ((strtotime($id[0]["executed"]) <= 0) OR ($id[0]["pid"] == 0)) {
-			logger("Entry for queue item ".$r[0]["id"]." wasn't stored - skip this execution", LOGGER_DEBUG);
+		logger("Entry for queue item ".$queue["id"]." wasn't stored - skip this execution", LOGGER_DEBUG);
 		q("COMMIT");
-			continue;
+		return true;
 	} elseif ($id[0]["pid"] != $mypid) {
-			logger("Queue item ".$r[0]["id"]." is to be executed by process ".$id[0]["pid"]." and not by me (".$mypid.") - skip this execution", LOGGER_DEBUG);
+		logger("Queue item ".$queue["id"]." is to be executed by process ".$id[0]["pid"]." and not by me (".$mypid.") - skip this execution", LOGGER_DEBUG);
 		q("COMMIT");
-			continue;
+		return true;
 	}
 	q("COMMIT");
 
-		$argv = json_decode($r[0]["parameter"]);
+	$argv = json_decode($queue["parameter"]);
 
 	$argc = count($argv);
 
@@ -128,8 +147,8 @@ function poller_run(&$argv, &$argc){
 
 	if (!validate_include($include)) {
 		logger("Include file ".$argv[0]." is not valid!");
-			q("DELETE FROM `workerqueue` WHERE `id` = %d", intval($r[0]["id"]));
-			continue;
+		q("DELETE FROM `workerqueue` WHERE `id` = %d", intval($queue["id"]));
+		return true;
 	}
 
 	require_once($include);
@@ -137,7 +156,7 @@ function poller_run(&$argv, &$argc){
 	$funcname = str_replace(".php", "", basename($argv[0]))."_run";
 
 	if (function_exists($funcname)) {
-			logger("Process ".$mypid." - Prio ".$r[0]["priority"]." - ID ".$r[0]["id"].": ".$funcname." ".$r[0]["parameter"]);
+		logger("Process ".$mypid." - Prio ".$queue["priority"]." - ID ".$queue["id"].": ".$funcname." ".$queue["parameter"]);
 
 		// For better logging create a new process id for every worker call
 		// But preserve the old one for the worker
@@ -149,21 +168,18 @@ function poller_run(&$argv, &$argc){
 		$a->process_id = $old_process_id;
 
 		if ($cooldown > 0) {
-				logger("Process ".$mypid." - Prio ".$r[0]["priority"]." - ID ".$r[0]["id"].": ".$funcname." - in cooldown for ".$cooldown." seconds");
+			logger("Process ".$mypid." - Prio ".$queue["priority"]." - ID ".$queue["id"].": ".$funcname." - in cooldown for ".$cooldown." seconds");
 			sleep($cooldown);
 		}
 
-			logger("Process ".$mypid." - Prio ".$r[0]["priority"]." - ID ".$r[0]["id"].": ".$funcname." - done");
+		logger("Process ".$mypid." - Prio ".$queue["priority"]." - ID ".$queue["id"].": ".$funcname." - done");
 
-			q("DELETE FROM `workerqueue` WHERE `id` = %d", intval($r[0]["id"]));
-		} else
+		q("DELETE FROM `workerqueue` WHERE `id` = %d", intval($queue["id"]));
+	} else {
 		logger("Function ".$funcname." does not exist");
-
-		// Quit the poller once every hour
-		if (time() > ($starttime + 3600))
-			return;
 	}
 
+	return true;
 }
 
 /**
@@ -295,6 +311,11 @@ function poller_kill_stale_workers() {
 		}
 }
 
+/**
+ * @brief Checks if the number of active workers exceeds the given limits
+ *
+ * @return bool Are there too much workers running?
+ */
 function poller_too_much_workers() {
 
 
@@ -373,6 +394,11 @@ function poller_too_much_workers() {
 	return($active >= $queues);
 }
 
+/**
+ * @brief Returns the number of active poller processes
+ *
+ * @return integer Number of active poller processes
+ */
 function poller_active_workers() {
 	$workers = q("SELECT COUNT(*) AS `processes` FROM `process` WHERE `command` = 'poller.php'");
 
@@ -394,8 +420,7 @@ function poller_passing_slow(&$highest_priority) {
 
 	$r = q("SELECT `priority`
 		FROM `process`
-		INNER JOIN `workerqueue` ON `workerqueue`.`pid` = `process`.`pid`
-		WHERE `process`.`command` = 'poller.php'");
+		INNER JOIN `workerqueue` ON `workerqueue`.`pid` = `process`.`pid`");
 
 	// No active processes at all? Fine
 	if (!dbm::is_result($r))
@@ -435,7 +460,6 @@ function poller_passing_slow(&$highest_priority) {
  *
  * @return string SQL statement
  */
-
 function poller_worker_process() {
 
 	q("START TRANSACTION;");
@@ -464,6 +488,62 @@ function poller_worker_process() {
 	return $r;
 }
 
+/**
+ * @brief Call the front end worker
+ */
+function call_worker() {
+	if (!get_config("system", "frontend_worker")) {
+		return;
+	}
+
+	$url = get_app()->get_baseurl()."/worker";
+	fetch_url($url, false, $redirects, 1);
+}
+
+/**
+ * @brief Call the front end worker if there aren't any active
+ */
+function call_worker_if_idle() {
+	if (!get_config("system", "frontend_worker")) {
+		return;
+	}
+
+	poller_run_cron();
+
+	clear_worker_processes();
+
+	$workers = q("SELECT COUNT(*) AS `processes` FROM `process` WHERE `command` = 'worker.php'");
+
+	if ($workers[0]["processes"] == 0) {
+		call_worker();
+	}
+}
+
+/**
+ * @brief Removes long running worker processes
+ */
+function clear_worker_processes() {
+	$timeout = Config::get("system", "frontend_worker_timeout", 10);
+
+	/// @todo We should clean up the corresponding workerqueue entries as well
+	q("DELETE FROM `process` WHERE `created` < '%s' AND `command` = 'worker.php'",
+		dbesc(datetime_convert('UTC','UTC',"now - ".$timeout." minutes")));
+}
+
+/**
+ * @brief Runs the cron processes
+ */
+function poller_run_cron() {
+	// Run the cron job that calls all other jobs
+	proc_run(PRIORITY_MEDIUM, "include/cron.php");
+
+	// Run the cronhooks job separately from cron for being able to use a different timing
+	proc_run(PRIORITY_MEDIUM, "include/cronhooks.php");
+
+	// Cleaning dead processes
+	poller_kill_stale_workers();
+}
+
 if (array_search(__file__,get_included_files())===0){
 	poller_run($_SERVER["argv"],$_SERVER["argc"]);
 

index.php

@@ -99,6 +99,10 @@ if (!$a->is_backend()) {
 	$stamp1 = microtime(true);
 	session_start();
 	$a->save_timestamp($stamp1, "parser");
+} else {
+	require_once "include/poller.php";
+
+	call_worker_if_idle();
 }
 
 /**

mod/worker.php

@@ -0,0 +1,47 @@
+<?php
+/**
+ * @file mod/worker.php
+ * @brief Module for running the poller as frontend process
+ */
+require_once("include/poller.php");
+
+use \Friendica\Core\Config;
+use \Friendica\Core\PConfig;
+
+function worker_init($a){
+
+	if (!Config::get("system", "frontend_worker")) {
+		return;
+	}
+
+	clear_worker_processes();
+
+	$workers = q("SELECT COUNT(*) AS `processes` FROM `process` WHERE `command` = 'worker.php'");
+
+	if ($workers[0]["processes"] > Config::get("system", "worker_queues", 4)) {
+		return;
+	}
+
+	$a->start_process();
+
+	logger("Front end worker started: ".getmypid());
+
+	call_worker();
+
+	if ($r = poller_worker_process()) {
+
+		// On most configurations this parameter wouldn't have any effect.
+		// But since it doesn't destroy anything, we just try to get more execution time in any way.
+		set_time_limit(0);
+
+		poller_execute($r[0]);
+	}
+
+	call_worker();
+
+	$a->end_process();
+
+	logger("Front end worker ended: ".getmypid());
+
+	killme();
+}

util/daemon.php

@@ -0,0 +1,101 @@
+<?php
+/**
+ * @file util/daemon.php
+ * @brief Run the poller from a daemon.
+ *
+ * This script was taken from http://php.net/manual/en/function.pcntl-fork.php
+ */
+function shutdown() {
+	posix_kill(posix_getpid(), SIGHUP);
+}
+
+if (in_array("start", $_SERVER["argv"])) {
+	$mode = "start";
+}
+
+if (in_array("stop", $_SERVER["argv"])) {
+	$mode = "stop";
+}
+
+if (in_array("status", $_SERVER["argv"])) {
+	$mode = "status";
+}
+
+if (!isset($mode)) {
+	die("Please use either 'start', 'stop' or 'status'.\n");
+}
+
+@include(".htconfig.php");
+
+if (!isset($pidfile)) {
+	die('Please specify a pid file in the variable $pidfile in the .htconfig.php. For example:'."\n".
+		'$pidfile = "/path/to/daemon.pid";'."\n");
+}
+
+if (in_array($mode, array("stop", "status"))) {
+	$pid = @file_get_contents($pidfile);
+
+	if (!$pid) {
+		die("Pidfile wasn't found. Is the daemon running?\n");
+	}
+}
+
+if ($mode == "status") {
+	if (posix_kill($pid, 0)) {
+		die("Daemon process $pid is running.\n");
+	}
+
+	unlink($pidfile);
+
+	die("Daemon process $pid isn't running.\n");
+}
+
+if ($mode == "stop") {
+	posix_kill($pid, SIGTERM);
+
+	unlink($pidfile);
+
+	die("Worker daemon process $pid was killed.\n");
+}
+
+echo "Starting worker daemon.\n";
+
+if (isset($a->config['php_path'])) {
+	$php = $a->config['php_path'];
+} else {
+	$php = "php";
+}
+
+// Switch over to daemon mode.
+if ($pid = pcntl_fork())
+	return;     // Parent
+
+fclose(STDIN);  // Close all of the standard
+fclose(STDOUT); // file descriptors as we
+fclose(STDERR); // are running as a daemon.
+
+register_shutdown_function('shutdown');
+
+if (posix_setsid() < 0)
+	return;
+
+if ($pid = pcntl_fork())
+	return;     // Parent
+
+$pid = getmypid();
+file_put_contents($pidfile, $pid);
+
+// Now running as a daemon.
+while (true) {
+	// Just to be sure that this script really runs endlessly
+	set_time_limit(0);
+
+	// Call the poller
+	$cmdline = $php.' include/poller.php';
+
+	exec($cmdline);
+
+	// Now sleep for 5 minutes
+	sleep(300);
+}
+?>