Google App Engine Memcache跨线程调用Set报错，求解决方案

Hey there, let's work through this Memcache problem you're hitting on Google App Engine! That error message you're seeing—Can't make API call memcache.Set in a thread that is neither the original request thread nor a thread created by ThreadManager—comes from GAE's strict thread context rules for its native Memcache API. The API is tied to the request's context, so it only works in threads that are either the main request thread or ones explicitly created via GAE's ThreadManager.

Here are three solid solutions to get around this:

1. Use GAE's ThreadManager for Background Threads #

If you absolutely need to run Memcache operations in a separate thread, create that thread using ThreadManager instead of a regular Java Thread. This ensures the thread inherits the request context, making Memcache calls valid. Here's a quick example:

import com.google.appengine.api.ThreadManager;
import com.google.appengine.api.memcache.MemcacheService;
import com.google.appengine.api.memcache.MemcacheServiceFactory;

// Inside your request-handling code (like a Servlet's doGet method)
Runnable memcacheTask = () -> {
    MemcacheService memcache = MemcacheServiceFactory.getMemcacheService();
    memcache.put("user:123", "John Doe"); // Safe to call here!
};

// Create a thread bound to the current request context
Thread backgroundThread = ThreadManager.createThreadForCurrentRequest(memcacheTask);
backgroundThread.start();
backgroundThread.join(); // Wait for the task to finish if needed

2. Keep Memcache Operations in the Original Request Thread #

The simplest fix is to avoid spawning custom threads altogether. If your use case allows, run all Memcache reads/writes directly in the main thread that handles the incoming request (like your Servlet's doGet/doPost methods or your endpoint handler). This bypasses the thread restriction entirely since you're using the context-bound request thread.

3. Switch to Cloud Memorystore (Redis) #

If you need more flexibility with threading (or want extra features like data persistence, clustering, or advanced caching patterns), consider migrating to Cloud Memorystore for Redis. It's a fully managed Redis service on Google Cloud that doesn't have GAE's native Memcache thread restrictions—you can call it from any thread in your app.

Here's a quick snippet using the Jedis client to interact with Memorystore:

import redis.clients.jedis.Jedis;

public class RedisDemo {
    public static void main(String[] args) {
        // Replace with your Memorystore host and port
        try (Jedis jedis = new Jedis("your-memorystore-host", 6379)) {
            jedis.set("product:456", "Wireless Headphones");
            String productName = jedis.get("product:456");
            System.out.println(productName); // Outputs "Wireless Headphones"
        }
    }
}

Just note that you'll need to enable the Memorystore service in your Google Cloud project and configure the correct network access (like allowing your GAE app to connect to the Redis instance).

内容的提问来源于stack exchange，提问作者SkyWalker