Guava本地缓存托底缓存以及异步更新缓存
1.简介
- 1.1 guava本地缓存是开发中比较常用的组件，一般使用 LoadingCache，将需要的值加载在内存中，如下所示
LoadingCache<String,T> cacheLoader= CacheBuilder
.newBuilder()
.expireAfterWrite(5, TimeUnit.MINUTES)
.build(new CacheLoader<String, T>() {
@Override
public T load(String key) throws Exception {
return method(key);//1.执行method,获取key对应的值
}
});
使用的方法:
T value=cacheLoader.get(key);//获取key对应的值
2.托底缓存设置
- 如果mehod()执行出错的话，无法拿到新的缓存。有时候，我们希望如果method执行异常的时候，本地缓存依旧用过期的缓存，那么可以重写CacheLoader中的reload方法进行设置
public abstract class RefreshKeepCacheLoader<K, V> extends CacheLoader<K, V> {
public ListenableFuture<V> reload(K key, V oldValue) throws Exception {
Object newvalue = null;
try {
newvalue = this.load(key);
} catch (Exception e) {
}
if(newvalue == null) {
newvalue = oldValue;
}
return Futures.immediateFuture(newvalue);
}
}
那么此时我们的cacheLoader可以这么写：
LoadingCache<String,T> cacheLoader= CacheBuilder
.newBuilder()
.expireAfterWrite(5, TimeUnit.MINUTES)
.build(new RefreshKeepCacheLoader<String, T>() {
@Override
public T load(String key) throws Exception {
return method(key);//2.执行method,获取key对应的值
}
});
与上面不同的是，用自定义的RefreshKeepCacheLoader替换了CacheLoader类，由于缓存过期会执行reload方法，如果reload异常，就采用oldValue。
3.异步缓存设置
- 3.1在并发条件下，有N个线程，如果缓存失效了，会有一个线程A去执行load方法（参见官方说明文档If another call to get(K) or getUnchecked(K) is currently loading the value for key, simply waits for that thread to finish and returns its loaded value.）,而其他线程就会等待线程A得到的结果，这样就会影响性能。
- 3.2 我们可以使全部线程返回旧的缓存值，把去异步更新缓存。方法如下：
首先我们看下，默认的reload的方法是怎么写的:
public ListenableFuture<V> reload(K key, V oldValue) throws Exception {
checkNotNull(key);
checkNotNull(oldValue);
return Futures.immediateFuture(load(key));
}
其实通过方法名，就可以看出执行load方法，然后用Futures.immediateFuture封装成ListenableFuture，再来看下immediateFuture方法
/**
* Creates a {@code ListenableFuture} which has its value set immediately upon construction. The
* getters just return the value. This {@code Future} can't be canceled or timed out and its
* {@code isDone()} method always returns {@code true}.
*/
public static <V> ListenableFuture<V> immediateFuture(@Nullable V value) {
if (value == null) {
// This cast is safe because null is assignable to V for all V (i.e. it is covariant)
@SuppressWarnings({"unchecked", "rawtypes"})
ListenableFuture<V> typedNull = (ListenableFuture) ImmediateSuccessfulFuture.NULL;
return typedNull;
}
return new ImmediateSuccessfulFuture<V>(value);
}
这就比较清楚了，这个future中直接设置的是值。
那么我们现在就要重写这个reload方法。
public abstract class RefreshAsyncCacheLoader<K, V> extends CacheLoader<K, V> {
@Override
public ListenableFuture<V> reload(final K key, final V oldValue) throws Exception {
ListenableFutureTask<V> task = ListenableFutureTask.create(new Callable<V>() {
public V call() {
try {
return load((K) key);
} catch (Exception e) {
}
return oldValue;
}
});
ThreadPoolUtil.getInstance().execute(task);//这里将这个task放到自定义的线程池中去执行，返回一个futrue，可以通过future获取线程执行获取的值
return task;
}
}
4.总结
- 最终我们可以得到具有托底缓存设置，并且可以异步更新缓存的guavaCache
LoadingCache<String,T> cacheLoader= CacheBuilder
.newBuilder()
.expireAfterWrite(5, TimeUnit.MINUTES)
.build(new RefreshAsyncCacheLoader <String, T>() {
@Override
public T load(String key) throws Exception {
return method(key);//1.执行method,获取key对应的值
}
});
5.实验
下面我们做一个简单的实验，涉及到WebController，CacheTest这两个类。
package com.netease.mail.activity.web.controller;
import com.netease.mail.activity.constant.RetCode;
import com.netease.mail.activity.meta.RequestHolder;
import com.netease.mail.activity.meta.vo.common.AjaxResult;
import com.netease.mail.activity.service.CacheTest;
import com.netease.mail.rambo.log.StatLogger;
import com.netease.mail.rambo.log.StatLoggerFactory;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Controller;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RequestMethod;
import org.springframework.web.bind.annotation.ResponseBody;
import javax.servlet.http.HttpServletRequest;
/**
* 重构web-control
* Created by hzlaojiaqi on 2016/9/19.
*/
@Controller
public class WebController {
private static final Logger INTERACTIVE_LOG = LoggerFactory.getLogger("INTERACTIVE_LOGGER");
public static final StatLogger USER_TRACE_LOG = StatLoggerFactory.getLogger("emptyProject");
@Autowired
RequestHolder mRequestHolder;
@Autowired
CacheTest cacheTest;
/**
* 获取当前uid的信息
* @param httpServletRequest
* @return
*/
@RequestMapping(value = "/ajax/getActInfo.do",method = RequestMethod.GET)
@ResponseBody
public AjaxResult getActInfo2(HttpServletRequest httpServletRequest){
String uid = mRequestHolder.getUid();
return new AjaxResult(RetCode.SUCCESS,cacheTest.cacheGet(uid));
}
}
- 在看下CacheTest这个service
package com.netease.mail.activity.service;
import com.netease.mail.activity.cache.RedisConfigure;
import com.netease.mail.activity.web.controller.BaseAjaxController;
import com.netease.mail.util.common.TimeUtil;
import net.sf.ehcache.Cache;
import net.sf.ehcache.CacheManager;
import net.sf.ehcache.Element;
import org.springframework.cache.annotation.CachePut;
import org.springframework.cache.annotation.Cacheable;
import org.springframework.stereotype.Service;
/**
* Created by hzlaojiaqi on 2017/11/9.
*/
@Service
public class CacheTest extends BaseService{
@Cacheable(cacheManager = "activityCacheManager",key = "'actCache_'+#uid",cacheNames = "actCache",unless = "#result==null")
public String cacheGet(String uid){
String s = uid + "_" + TimeUtil.now();
THIRDPARTY_LOG.info("cacheGet in method uid:{} s:{}",uid,s);
return s;
}
}
- 这里设置缓存的时间比较短，只有10s钟，具体的设置如下：
LoadingCache<String,Object> cache = CacheBuilder
.newBuilder()
.maximumSize(5210)
.refreshAfterWrite(10, TimeUnit.SECONDS)
.build(new RefreshAsyncCacheLoader<String, Object>() {
@Override
public Object load(String key) throws Exception {
Object o = mRedis.opsForValue().get(key);
THIRDPARTY_LOG.debug("thread:{},activityManager local cache reload key:{},result:{}",Thread.currentThread().getName(),key,o);
return o;
}
});
- 最终可以看到的log为
- 请求线程为http-apr-8080-exec-8，该线程马上返回了原来的值，而我们自定义的线程custom_thread_member_11则执行了reload方法。（这里加载后的值和之前的值是一样的，特此说明）
- 如果是第一次请求，由于没有旧的值，那么http-apr-8080-exec-8会去执行reload方法。
Ps：文章难免有纰漏，望拍砖指正，感谢。