otherwise you'll get a little bit of spam when workers die instantly
@@ -12,7 +12,7 @@ if (cluster.isMaster) {
cluster.on("exit", function (worker, code, signal) {
logging.error("Worker died. Rebooting a new one.");
- cluster.fork();
+ setTimeout(cluster.fork, 100);
});
setInterval(cleaner.run, config.cleaning_interval * 1000);