作者:david.turing 来源:BlogJava 酷勤网收集 2008-04-20
最近协助一些BEA客户做调优,他们使用了Spring,出现了各种各样的性能问题,这些问题其实都是不容易重现的,其中,我自己捕获了一些ThreadDump,并report了给Spring JIRA。这个Case的情况是:Spring会偶然出现CPU 100%的情况,WebLogic Server崩溃,我后来分析了线程Dump,觉得是一种Lock Contention的情形,幸好,Juergen Hoeller很快给我Fixed了这个Bug:
http://jira.springframework.org/browse/SPR-4664
使用Java编程的同学都建议Review一下,呵呵:
这是2.5.4以前的代码:
/**
* Cache of TransactionAttributes, keyed by DefaultCacheKey (Method + target Class).
* <p>As this base class is not marked Serializable, the cache will be recreated
* after serialization - provided that the concrete subclass is Serializable.
*/
final Map attributeCache = new HashMap();
/**
* Determine the transaction attribute for this method invocation.
* <p>Defaults to the class's transaction attribute if no method attribute is found.
* @param method the method for the current invocation (never <code>null</code>)
* @param targetClass the target class for this invocation (may be <code>null</code>)
* @return TransactionAttribute for this method, or <code>null</code> if the method
* is not transactional
*/
public TransactionAttribute getTransactionAttribute(Method method, Class targetClass) {
// First, see if we have a cached value.
Object cacheKey = getCacheKey(method, targetClass);
synchronized ( this .attributeCache) {
Object cached = this .attributeCache.get(cacheKey);
if (cached != null ) {
// Value will either be canonical value indicating there is no transaction attribute,
// or an actual transaction attribute.
if (cached == NULL_TRANSACTION_ATTRIBUTE) {
return null ;
}
else {
return (TransactionAttribute) cached;
}
}
else {
// We need to work it out.
TransactionAttribute txAtt = computeTransactionAttribute(method, targetClass);
// Put it in the cache.
if (txAtt == null ) {
this .attributeCache.put(cacheKey, NULL_TRANSACTION_ATTRIBUTE);
}
else {
if (logger.isDebugEnabled()) {
logger.debug( " Adding transactional method [ " + method.getName() + " ] with attribute [ " + txAtt + " ] " );
}
this .attributeCache.put(cacheKey, txAtt);
}
return txAtt;
}
}
}
这是2.5.4 Fixed后的代码:
/**
* Cache of TransactionAttributes, keyed by DefaultCacheKey (Method + target Class).
* <p>As this base class is not marked Serializable, the cache will be recreated
* after serialization - provided that the concrete subclass is Serializable.
*/
final Map attributeCache = CollectionFactory.createConcurrentMapIfPossible( 16 );
/**
* Determine the transaction attribute for this method invocation.
* <p>Defaults to the class's transaction attribute if no method attribute is found.
* @param method the method for the current invocation (never <code>null</code>)
* @param targetClass the target class for this invocation (may be <code>null</code>)
* @return TransactionAttribute for this method, or <code>null</code> if the method
* is not transactional
*/
public TransactionAttribute getTransactionAttribute(Method method, Class targetClass) {
// First, see if we have a cached value.
Object cacheKey = getCacheKey(method, targetClass);
Object cached = this .attributeCache.get(cacheKey);
if (cached != null ) {
// Value will either be canonical value indicating there is no transaction attribute,
// or an actual transaction attribute.
if (cached == NULL_TRANSACTION_ATTRIBUTE) {
return null ;
}
else {
return (TransactionAttribute) cached;
}
}
else {
// We need to work it out.
TransactionAttribute txAtt = computeTransactionAttribute(method, targetClass);
// Put it in the cache.
if (txAtt == null ) {
this .attributeCache.put(cacheKey, NULL_TRANSACTION_ATTRIBUTE);
}
else {
if (logger.isDebugEnabled()) {
logger.debug( " Adding transactional method [ " + method.getName() + " ] with attribute [ " + txAtt + " ] " );
}
this .attributeCache.put(cacheKey, txAtt);
}
return txAtt;
}
}
但是2.5.4 snapshot是未经很好测试的版本,客户一般不太敢用。我不知道其实有多少客户真正地把Spring投入到高并发性环境下使用,如果有,他们应该会能碰到我所碰到的情形。
-
# re: 将Spring用于高并发环境的隐忧[未登录] BeanSoft Posted @ 2008-04-19 10:03
没错,我们以前的公司也是Weblogic + Hibernate,出了性能问题了,虽然调整解决了,但不管怎么说,因为这些开源软件之前开发的时候并没有考虑高并发和集群的情况,还是比较容易出现问题的,尤其是没有经过严格的压力测试。我个人认为,目前做的比较好的软件,依然是商业的。 -
-
# re: 将Spring用于高并发环境的隐忧 flyisland Posted @ 2008-04-19 10:56
1. 的确,许多开源软件在版本发布之前是否有能力去做高并发压力测试,值得怀疑
2. 我想,看到david这次发现的问题,用户更应该当心的是下次如果某个开源产品出了问题怎么办。
3. 不过目前来说是,不是“用不用开源的问题”(可以是一定要用),而是“用了开源怎么办”的问题。所以这是开源商业模式的一个机会,比如david就可以去做开源调优的服务支持了,哈
4. 这篇blog在我的firefox 3.0b5下排版混乱,不得不到IE下面来留言 -
-
# re: 将Spring用于高并发环境的隐忧[未登录] af Posted @ 2008-04-19 17:43
我一直认为开源代码是用来学习的,实际项目中还是不要用的好。 -
-
# re: 将Spring用于高并发环境的隐忧[未登录] lj Posted @ 2008-04-19 21:08
我很想知道在spring1.2.9或spring2.0.8有没有这个问题 -
# re: 将Spring用于高并发环境的隐忧 GoGo Posted @ 2008-04-20 13:24
@flyisland
FireFox不仅仅3.0乱,2.0也乱,于是IE留言之。
来自:http://www.blogjava.net/security/archive/2008/04/19/spring_bug.html

