GL shutdown on catchable OutOfMemoryError (#15889)

* GL shutdown on catchable OutOfMemoryError

* Changelog added

* Changelog renamed, because of mixed-repo situation

* New entry in UPGRADING.md

UPGRADING.md

@@ -12,6 +12,11 @@ Upgrading to Graylog 5.2.x
 - GreyNoise Full IP Lookup [Enterprise] Data Adapter can no longer be used with a free GreyNoise Community API tokens.
 - GreyNoise Quick IP Lookup Data Adapter can no longer be used with a free GreyNoise Community API tokens.
 
+## Shutdown of Graylog on OutOfMemoryError
+Because of an error in HttpCore 4.4.12, which is required by Elasticsearch and older versions of Opensearch, OutOfMemoryError errors were not properly handled.
+The Reactor was stopped, which prevented proper Graylog operation and the reason (OutOfMemoryError) was not clearly visible.
+From now on, Graylog will shutdown on OutOfMemoryError, trying to log some basic information about the thread and memory consumption during this event.
+
 ## Java API Changes
 The following Java Code API changes have been made.
 

changelog/unreleased/pr-15889.toml

@@ -0,0 +1,5 @@
+type = "fixed" # One of: a(dded), c(hanged), d(eprecated), r(emoved), f(ixed), s(ecurity)
+message = "Fix problem with STOPPED reactor on OutOfMemoryError. This error stops the whole GL instance now, instead of constantly logging errors related to stopped reactor."
+
+issues = ["graylog-plugin-enterprise#4980"]
+pulls = ["15889"]

graylog2-server/src/main/java/org/graylog2/bindings/ServerBindings.java

@@ -36,6 +36,7 @@
 import org.graylog2.bindings.providers.SecureFreemarkerConfigProvider;
 import org.graylog2.bindings.providers.SystemJobFactoryProvider;
 import org.graylog2.bindings.providers.SystemJobManagerProvider;
+import org.graylog2.bootstrap.uncaughtexeptions.DefaultUncaughtExceptionHandlerCreator;
 import org.graylog2.cluster.ClusterConfigServiceImpl;
 import org.graylog2.cluster.leader.FakeLeaderElectionModule;
 import org.graylog2.cluster.leader.LeaderElectionModule;
@@ -178,6 +179,7 @@ private void bindFactoryModules() {
     }
 
     private void bindSingletons() {
+        bind(DefaultUncaughtExceptionHandlerCreator.class).asEagerSingleton();
         bind(SystemJobManager.class).toProvider(SystemJobManagerProvider.class);
         bind(DefaultSecurityManager.class).toProvider(DefaultSecurityManagerProvider.class).asEagerSingleton();
         bind(SystemJobFactory.class).toProvider(SystemJobFactoryProvider.class);

graylog2-server/src/main/java/org/graylog2/bootstrap/uncaughtexeptions/DefaultUncaughtExceptionHandlerCreator.java

@@ -0,0 +1,56 @@
+/*
+ * Copyright (C) 2020 Graylog, Inc.
+ *
+ * This program is free software: you can redistribute it and/or modify
+ * it under the terms of the Server Side Public License, version 1,
+ * as published by MongoDB, Inc.
+ *
+ * This program 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
+ * Server Side Public License for more details.
+ *
+ * You should have received a copy of the Server Side Public License
+ * along with this program. If not, see
+ * <http://www.mongodb.com/licensing/server-side-public-license>.
+ */
+package org.graylog2.bootstrap.uncaughtexeptions;
+
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import javax.inject.Singleton;
+
+@Singleton
+public class DefaultUncaughtExceptionHandlerCreator {
+
+    private static final Logger LOG = LoggerFactory.getLogger(DefaultUncaughtExceptionHandlerCreator.class);
+
+    public DefaultUncaughtExceptionHandlerCreator() {
+        Thread.setDefaultUncaughtExceptionHandler((t, e) -> {
+            if (!(e instanceof ThreadDeath)) {
+                defaultHandling(t, e);
+                if (e instanceof OutOfMemoryError) {
+                    outOfMemoryHandling(t);
+                }
+            }
+        });
+    }
+
+    private void outOfMemoryHandling(final Thread t) {
+        LOG.error("OutOfMemoryError encountered in thread " + t.getName() + ", Graylog instance will be shut down.");
+        final Runtime runtime = Runtime.getRuntime();
+        LOG.info("Free JVM memory : " + runtime.freeMemory());
+        LOG.info("Total JVM memory : " + runtime.totalMemory());
+        LOG.info("Max JVM memory : " + runtime.maxMemory());
+
+        System.exit(1);
+    }
+
+    private void defaultHandling(final Thread t, final Throwable e) {
+        //see ThreadGrooup.uncaughtException -> we don't want to remove the well-known "printStackTrace" uncaught exception handling
+        System.err.print("Exception in thread \""
+                + t.getName() + "\" ");
+        e.printStackTrace(System.err);
+    }
+}