The Tomcat support community replied with the following. I'm thinking anyone using BlazeDS with Tomcat should see the same issue. Anyone know how to resolve?
-------tomcat comments follow---------- If all goes well, this leak will be handled by Tomcat and your service won't suffer for it. That said, you should fix the problem because cleaning-up messes is wasteful if the mess wasn't necessary in the first place. The leak itself is coming from your application or one of the libraries it's using. The solution will be to find and fix that leak. I would start by asking the Apache Flex people what the SerializationContext is for, and how to remove ThreadLocal values from shared threads (such as those in a servlet environment). It's possible you are using the Flex framework in a way that is not conducive to a servlet environment, but that a few changes could make it safer to use. ----- Original Message ----- From: modjkl...@comcast.net To: "apache users" <users@flex.apache.org> Sent: Friday, January 6, 2017 2:12:59 PM Subject: Re: Tomcat 8 memory leak (?) moving from GlassFish I should mention I'm using BlazeDS, which I'm assuming is creating connections(?). ----- Original Message ----- From: modjkl...@comcast.net To: "apache flex users" <users@flex.apache.org> Sent: Friday, January 6, 2017 2:08:08 PM Subject: Tomcat 8 memory leak (?) moving from GlassFish In the processing of moving my Flex 4.12 web app from GlassFish 3.1.2 to Tomcat 8.5.9, I observe the following SEVERE error in the tomcat log. Has anyone seen this before, and if so, recommend how to address it? 06-Jan-2017 13:49:07.644 SEVERE [ContainerBackgroundProcessor[StandardEngine[Catalina]]] org.apache.catalina.loader.WebappClassLoaderBase.checkThreadLocalMapForLeaks The web application [myApp] created a ThreadLocal with key of type [java.lang.ThreadLocal] (value [java.lang.ThreadLocal@4d6e6d7b]) and a value of type [flex.messaging.io.SerializationContext] (value [flex.messaging.io.SerializationContext@66135428]) but failed to remove it when the web application was stopped. Threads are going to be renewed over time to try and avoid a probable memory leak.