Double check for maximum number of workers

[friendica.git] / include / poller.php

diff --git a/include/poller.php b/include/poller.php
index e33167c5b14667eba38e01b18a38e5da755ccbe5..3b348531c5b6ed47806b1fa821189bb7d0262b17 100644 (file)
--- a/include/poller.php
+++ b/include/poller.php
@@ -26,10 +26,11 @@ function poller_run(&$argv, &$argc){
                unset($db_host, $db_user, $db_pass, $db_data);
        };
 
-       $maxsysload = intval(get_config('system','maxloadavg'));
-       if($maxsysload < 1)
-               $maxsysload = 50;
        if(function_exists('sys_getloadavg')) {
+               $maxsysload = intval(get_config('system','maxloadavg'));
+               if($maxsysload < 1)
+                       $maxsysload = 50;
+
                $load = sys_getloadavg();
                if(intval($load[0]) > $maxsysload) {
                        logger('system: load ' . $load[0] . ' too high. poller deferred to next scheduled run.');
@@ -50,39 +51,42 @@ function poller_run(&$argv, &$argc){
                        if (!posix_kill($pid["pid"], 0))
                                q("UPDATE `workerqueue` SET `executed` = '0000-00-00 00:00:00', `pid` = 0 WHERE `pid` = %d",
                                        intval($pid["pid"]));
+                       else {
+                               // To-Do: Kill long running processes
+                               // But: Update processes (like the database update) mustn't be killed
+                       }
+
+       } else {
+               // Checking the number of workers
+               if (poller_too_much_workers(1))
+                       return;
 
-       } else
-               // Sleep two seconds before checking for running processes to avoid having too many workers
+               // Sleep four seconds before checking for running processes again to avoid having too many workers
                sleep(4);
+       }
 
        // Checking number of workers
-       if (poller_too_much_workers())
+       if (poller_too_much_workers(2))
                return;
 
        $starttime = time();
 
        while ($r = q("SELECT * FROM `workerqueue` WHERE `executed` = '0000-00-00 00:00:00' ORDER BY `created` LIMIT 1")) {
 
-               if(function_exists('sys_getloadavg')) {
-                       $load = sys_getloadavg();
-                       if(intval($load[0]) > $maxsysload) {
-                               logger('system: load ' . $load[0] . ' too high. poller deferred to next scheduled run.');
-                               return;
-                       }
-               }
-
-               // Quit the poller once every hour
-               if (time() > ($starttime + 3600))
-                       return;
-
-               if (poller_too_much_workers())
-                       return;
-
-               q("UPDATE `workerqueue` SET `executed` = '%s', `pid` = %d WHERE `id` = %d",
+               q("UPDATE `workerqueue` SET `executed` = '%s', `pid` = %d WHERE `id` = %d AND `executed` = '0000-00-00 00:00:00'",
                        dbesc(datetime_convert()),
                        intval(getmypid()),
                        intval($r[0]["id"]));
 
+               // Assure that there are no tasks executed twice
+               $id = q("SELECT `id` FROM `workerqueue` WHERE `id` = %d AND `pid` = %d",
+                       intval($r[0]["id"]),
+                       intval(getmypid()));
+               if (!$id) {
+                       logger("Queue item ".$r[0]["id"]." was executed multiple times - skip this execution", LOGGER_DEBUG);
+                       continue;
+               }
+
                $argv = json_decode($r[0]["parameter"]);
 
                $argc = count($argv);
@@ -109,34 +113,47 @@ function poller_run(&$argv, &$argc){
                        q("DELETE FROM `workerqueue` WHERE `id` = %d", intval($r[0]["id"]));
                } else
                        logger("Function ".$funcname." does not exist");
+
+               // Quit the poller once every hour
+               if (time() > ($starttime + 3600))
+                       return;
+
+               // Count active workers and compare them with a maximum value that depends on the load
+               if (poller_too_much_workers(3))
+                       return;
        }
 
 }
 
-function poller_too_much_workers() {
+function poller_too_much_workers($stage) {
 
+       $queues = get_config("system", "worker_queues");
+
+       if ($queues == 0)
+               $queues = 4;
+
+       $active = poller_active_workers();
+
+       // Decrease the number of workers at higher load
        if(function_exists('sys_getloadavg')) {
-               $load = sys_getloadavg();
+               $load = max(sys_getloadavg());
 
-               // To-Do
-               if ($load < 1)
-                       $queues = 10;
-               elseif ($load < 5)
-                       $queues = 4;
-               elseif ($load < 10)
-                       $queues = 2;
-               else
-                       $queues = 1;
+               $maxsysload = intval(get_config('system','maxloadavg'));
+               if($maxsysload < 1)
+                       $maxsysload = 50;
 
-       } else {
-               $queues = intval(get_config("system", "worker_queues"));
+               $maxworkers = $queues;
+
+               // Some magical mathemathics to reduce the workers
+               $exponent = 3;
+               $slope = $maxworkers / pow($maxsysload, $exponent);
+               $queues = ceil($slope * pow(max(0, $maxsysload - $load), $exponent));
+
+               logger("Current load stage ".$stage.": ".$load." - maximum: ".$maxsysload." - current queues: ".$active." - maximum: ".$queues, LOGGER_DEBUG);
 
-               if ($queues == 0)
-                       $queues = 4;
        }
 
-       if (poller_active_workers() >= $queues)
-               return;
+       return($active >= $queues);
 }
 
 function poller_active_workers() {