当您希望某些状态不应该在不同线程之间共享，但每个线程在其整个生命周期内都可以访问时，ThreadLocal非常有用。

As an example, imagine a web application, where each request is served by a different thread. Imagine that for each request you need a piece of data multiple times, which is quite expensive to compute. However, that data might have changed for each incoming request, which means that you can't use a plain cache. A simple, quick solution to this problem would be to have a ThreadLocal variable holding access to this data, so that you have to calculate it only once for each request. Of course, this problem can also be solved without the use of ThreadLocal, but I devised it for illustration purposes.

也就是说，请记住ThreadLocals本质上是一种全局状态的形式。因此，它有许多其他含义，只有在考虑了所有其他可能的解决方案后才应该使用。

一种可能的(也是常见的)用途是，当您有一些不是线程安全的对象，但您希望避免同步访问该对象(我正在看您，SimpleDateFormat)。相反，给每个线程它自己的对象实例。

例如:

public class Foo
{
    // SimpleDateFormat is not thread-safe, so give one to each thread
    private static final ThreadLocal<SimpleDateFormat> formatter = new ThreadLocal<SimpleDateFormat>(){
        @Override
        protected SimpleDateFormat initialValue()
        {
            return new SimpleDateFormat("yyyyMMdd HHmm");
        }
    };

    public String formatIt(Date date)
    {
        return formatter.get().format(date);
    }
}

文档。

什么时候?

当一个对象不是线程安全的时候，为了避免影响可伸缩性的同步，给每个线程一个对象，并保持它的线程作用域，即ThreadLocal。最常用但不是线程安全的对象之一是数据库Connection和JMSConnection。

如何?

例如，Spring框架通过将这些连接对象保存在ThreadLocal变量中，在幕后大量使用ThreadLocal来管理事务。在高层，当事务启动时，它获得连接(并禁用自动提交)并将其保存在ThreadLocal中。在进一步的db调用中，它使用相同的连接与db通信。最后，它从ThreadLocal获取连接，提交(或回滚)事务并释放连接。

我认为log4j也使用ThreadLocal来维护MDC。

由于ThreadLocal是对给定线程中的数据的引用，因此在使用线程池的应用服务器中使用ThreadLocal时，可能会导致类加载泄漏。在使用ThreadLocal的remove()方法清理你get()或set()的任何ThreadLocals时，你需要非常小心。

如果你在完成时不清理，它持有的任何类的引用作为部署的webapp的一部分将保留在永久堆中，永远不会被垃圾收集。重新部署/取消部署webapp不会清除每个线程对你的webapp类的引用，因为线程不是你的webapp所拥有的。每次后续部署都将创建该类的新实例，该实例永远不会被垃圾收集。

由于java.lang.OutOfMemoryError: PermGen空间，你最终会出现内存不足的异常，在谷歌搜索之后可能只是增加-XX:MaxPermSize，而不是修复这个错误。

如果您最终遇到了这些问题，您可以通过使用Eclipse的Memory Analyzer和/或遵循Frank Kieviet的指南和后续内容来确定哪个线程和类保留了这些引用。

更新:重新发现Alex Vasseur的博客条目，它帮助我找到了一些我遇到的ThreadLocal问题。

试试这个小例子，感受一下ThreadLocal变量:

public class Book implements Runnable {
    private static final ThreadLocal<List<String>> WORDS = ThreadLocal.withInitial(ArrayList::new);

    private final String bookName; // It is also the thread's name
    private final List<String> words;


    public Book(String bookName, List<String> words) {
        this.bookName = bookName;
        this.words = Collections.unmodifiableList(words);
    }

    public void run() {
        WORDS.get().addAll(words);
        System.out.printf("Result %s: '%s'.%n", bookName, String.join(", ", WORDS.get()));
    }

    public static void main(String[] args) {
        Thread t1 = new Thread(new Book("BookA", Arrays.asList("wordA1", "wordA2", "wordA3")));
        Thread t2 = new Thread(new Book("BookB", Arrays.asList("wordB1", "wordB2")));
        t1.start();
        t2.start();
    }
}

控制台输出，如果线程BookA先执行: 结果BookA: 'wordA1, wordA2, wordA3'。 结果BookB: 'wordB1, wordB2'。 控制台输出，如果先执行线程BookB: 结果BookB: 'wordB1, wordB2'。 结果BookA: 'wordA1, wordA2, wordA3'。

在《Java并发实践》一书中有一个很好的例子。作者(Joshua Bloch)解释了线程限制是实现线程安全的最简单方法之一，而ThreadLocal是维护线程限制的更正式的方法。最后，他还解释了人们如何滥用它作为全局变量。

我已经从提到的书中复制了文本，但代码3.10是缺失的，因为它不太重要，了解ThreadLocal应该在哪里使用。

Thread-local variables are often used to prevent sharing in designs based on mutable Singletons or global variables. For example, a single-threaded application might maintain a global database connection that is initialized at startup to avoid having to pass a Connection to every method. Since JDBC connections may not be thread-safe, a multithreaded application that uses a global connection without additional coordination is not thread-safe either. By using a ThreadLocal to store the JDBC connection, as in ConnectionHolder in Listing 3.10, each thread will have its own connection. ThreadLocal is widely used in implementing application frameworks. For example, J2EE containers associate a transaction context with an executing thread for the duration of an EJB call. This is easily implemented using a static Thread-Local holding the transaction context: when framework code needs to determine what transaction is currently running, it fetches the transaction context from this ThreadLocal. This is convenient in that it reduces the need to pass execution context information into every method, but couples any code that uses this mechanism to the framework. It is easy to abuse ThreadLocal by treating its thread confinement property as a license to use global variables or as a means of creating “hidden” method arguments. Like global variables, thread-local variables can detract from reusability and introduce hidden couplings among classes, and should therefore be used with care.