http://lists.danga.com/pipermail/memcached/2007-July/004802.html
That's right. Memcached is a cache, not a primary data store. If you need to persist items to a database, you should do it incrementally as you populate the cache. In our environment (with a few exceptions) we only ever populate the cache with data we've read from a database -- that is, everything in the cache is guaranteed to be backed by a persistent store because it comes from a persistent store in the first place. When we update an object, we update it in the database and delete it from the cache, so the next reader that needs it can pull it from the database. That last part may or may not be optimal for your application (it essentially means the cache is lazy-populated) but the first part is a good guideline: if it's not already backed by a persistent store, don't put it in the cache. Follow that and you'll never need to worry about whether the cache can be dumped to disk. Dumping the cache to disk is dangerous in any event. Consider: User updates his account information Your code updates the database Your code updates the cache You dump the cache to disk and start working on the cache host User updates his account information again Your code updates the database Your code doesn't update the cache (because it's offline at the moment) You bring the cache back online and it restores its state User views his account information and gets the old data from the cache Obviously this race condition exists to some extent even without a persistent dump, but it's miniscule in comparison and is much easier to work around in the application code. On a busy site any persistent snapshot of the cache is pretty much guaranteed to be stale even before it's finished hitting the disk. That said, if you want a persistent store with a memcached interface, Google "Tugela cache" and you might find that more to your liking.
从缓存服务器中读取数据:
package com.memcached.test;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.TimeoutException;
import com.page.util.CurrentTime;
import net.rubyeye.xmemcached.KeyIterator;
import net.rubyeye.xmemcached.MemcachedClient;
import net.rubyeye.xmemcached.MemcachedClientBuilder;
import net.rubyeye.xmemcached.XMemcachedClientBuilder;
import net.rubyeye.xmemcached.exception.MemcachedException;
import net.rubyeye.xmemcached.utils.AddrUtil;
/**
* @description memcached的工具类,
* @author BrinPage
* @date 2012.07.19 19:12:00
*/
@SuppressWarnings("deprecation")
public class MemcachedTest{
/**
* 从缓存服务器上读取参数
* @param ip_port MemcachedServer的IP地址和端口号
* @return ArryList<String []>对象
*/
@SuppressWarnings("finally")
public List<String []> getMemcached(String ip_port){
List<String []> list = new ArrayList<String []>();
MemcachedClientBuilder builder = new XMemcachedClientBuilder(AddrUtil.getAddresses(ip_port));
builder.setConnectionPoolSize(5);
MemcachedClient memcachedClient = null;
CurrentTime time = new CurrentTime();
try {
memcachedClient = builder.build();
System.out.println(time.getCurrentTime());
KeyIterator it = memcachedClient.getKeyIterator(AddrUtil.getOneAddress(ip_port));
String key = null;
int count = 0;
while(it.hasNext())
{
key=it.next();
count ++;
String value [] = memcachedClient.get(key);
// System.out.println(value.length);
list.add(value);
}
System.out.println(time.getCurrentTime());
memcachedClient.shutdown();
} catch (MemcachedException e) {
e.printStackTrace();
} catch (InterruptedException e) {
e.printStackTrace();
} catch (TimeoutException e) {
e.printStackTrace();
} catch (IOException e) {
e.printStackTrace();
} finally{
return list;
}
}
public static void main(String args[]){
MemcachedTest test = new MemcachedTest();
List<String []> list = new ArrayList<String []>();
list = test.getMemcached("127.0.0.1:11211");
for(int i = 0; i < list.size(); i ++){
for(int j = 0; j < list.get(i).length; j ++){
System.out.print(list.get(i)[j] + "\t");
}
System.out.println();
System.out.println("*********************************************************");
}
}
}