[Hibernate-JIRA] Created: (HHH-5945) Race condition in building query cache
by Luca Domenichini (JIRA)
Race condition in building query cache
--------------------------------------
Key: HHH-5945
URL: http://opensource.atlassian.com/projects/hibernate/browse/HHH-5945
Project: Hibernate Core
Issue Type: Bug
Components: caching (L2)
Affects Versions: 3.6.1
Environment: Hibernate 3.6.1, EhCache 2.3.0..2.3.2, MSSQL 2005 Developer Edition
Reporter: Luca Domenichini
I have this sample program that creates 10 threads that executes the same cached query in a thread-local current session context. Entity ClasseAltezza is a READ_ONLY cached entity.
{panel}{noformat}
public static void main(String[] args) {
List<Thread> threads = new ArrayList<Thread>();
for (int i=0; i<10; i++) {
Thread t = new Thread() {
@Override
public void run() {
try {
HibernateUtil.beginTransaction();
HibernateUtil.getCurrentSession()
.createCriteria(ClasseAltezza.class)
.setCacheable(true)
.setCacheRegion(ClasseAltezza.class.toString())
.list();
HibernateUtil.commitTransaction();
}
catch (Exception ex) {
ex.printStackTrace();
HibernateUtil.rollbackTransaction();
}
}
};
threads.add(t);
}
for (Thread t : threads)
t.start();
}
{noformat}{panel}
This is the output:
{panel}{noformat}org.hibernate.cache.CacheException: net.sf.ehcache.ObjectExistsException: Cache class com.promag.wms.base.persistence.entities.ClasseAltezza already exists
at org.hibernate.cache.EhCacheProvider.buildCache(EhCacheProvider.java:101)
at org.hibernate.cache.impl.bridge.RegionFactoryCacheProviderBridge.buildQueryResultsRegion(RegionFactoryCacheProviderBridge.java:115)
at org.hibernate.cache.StandardQueryCache.<init>(StandardQueryCache.java:78)
at org.hibernate.cache.StandardQueryCacheFactory.getQueryCache(StandardQueryCacheFactory.java:44)
at org.hibernate.impl.SessionFactoryImpl.getQueryCache(SessionFactoryImpl.java:1173)
at org.hibernate.loader.Loader.listUsingQueryCache(Loader.java:2285)
at org.hibernate.loader.Loader.list(Loader.java:2268)
at org.hibernate.loader.criteria.CriteriaLoader.list(CriteriaLoader.java:119)
at org.hibernate.impl.SessionImpl.list(SessionImpl.java:1716)
at org.hibernate.impl.CriteriaImpl.list(CriteriaImpl.java:347)
at com.promag.wms.base.persistence.methods.ClasseAltezzaMethods.getAll(ClasseAltezzaMethods.java:30)
at Test$1.run(Test.java:20)
Caused by: net.sf.ehcache.ObjectExistsException: Cache class com.promag.wms.base.persistence.entities.ClasseAltezza already exists
at net.sf.ehcache.CacheManager.addCache(CacheManager.java:888)
at org.hibernate.cache.EhCacheProvider.buildCache(EhCacheProvider.java:94)
... 11 more
{noformat}{panel}
I see that in Hibernate 3.6.0 there were a synchronized block in method SessionFactory.getQueryCache(String regionName) the prevented the same query region to be built twice as it happens in Hibernate 3.6.1, were the synchronized block is missing..
This is the method in SessionFactoryImpl:
{panel}{noformat}
public QueryCache getQueryCache(String regionName) throws HibernateException {
if ( regionName == null ) {
return getQueryCache();
}
if ( !settings.isQueryCacheEnabled() ) {
return null;
}
synchronized ( allCacheRegions ) { // THIS SYNCHRONIZED IS MISSING IN 3.6.1!!!!
QueryCache currentQueryCache = ( QueryCache ) queryCaches.get( regionName );
if ( currentQueryCache == null ) {
currentQueryCache = settings.getQueryCacheFactory().getQueryCache( regionName, updateTimestampsCache, settings, properties );
queryCaches.put( regionName, currentQueryCache );
allCacheRegions.put( currentQueryCache.getRegion().getName(), currentQueryCache.getRegion() );
}
return currentQueryCache;
}
}
{noformat}{panel}
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators....
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
12 years, 6 months
[Hibernate-JIRA] Created: (HHH-2166) Long "in" lists in queries results in a Java stack overflow exception.
by Philip R. "Pib" Burns (JIRA)
Long "in" lists in queries results in a Java stack overflow exception.
----------------------------------------------------------------------
Key: HHH-2166
URL: http://opensource.atlassian.com/projects/hibernate/browse/HHH-2166
Project: Hibernate3
Type: Bug
Components: core
Versions: 3.2.0.ga
Environment: Hibernate 3.2.0.cr3 through 3.2.0.ga (at least). Any standard deployment of Sun's JVM on Windows, Linux, or Mac OS X (and presumably other platforms like Solaris)
Reporter: Philip R. "Pib" Burns
Attachments: NodeTraverser.java
With Hibernate 320ga a long "in" list can result in a stack overflow error during the parsing stage. For example, a query element like
where x in (:x)
or a manually constructed
where x in (1,2,3 .....)
can generate a stack overflow if the number of elements referenced by x exceeds a number dependent upon the amount of available stack space. For many JVMs, the limit is between 9,000 and 10,000 assuming a relatively empty stack at the point of query execution. We have applications which occasionally use lists several times this size.
The stack overflow occurs in org.hibernate.hql.ast.util.NodeTraverser which uses a recursive algorithm to walk a parse tree. Long "in" lists generate a subtree of depth about equal to the number of elements in the list. A sufficiently long list results in a stack overflow when NodeTraverser's internal method visitDepthFirst calls itself too many times.
The solution is to replace the recursive tree walking strategy with an iterative one that does not use up stack space. I am attaching the source for a replacement version of . NodeTraverser which implements the iterative tree walking method.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://opensource.atlassian.com/projects/hibernate/secure/Administrators....
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
12 years, 6 months
[Hibernate-JIRA] Created: (HHH-4577) 2L query cache: Low performance of flush and commit due many unnecessary (pre)invalidate calls on UpdateTimestampsCache
by Guenther Demetz (JIRA)
2L query cache: Low performance of flush and commit due many unnecessary (pre)invalidate calls on UpdateTimestampsCache
-----------------------------------------------------------------------------------------------------------------------
Key: HHH-4577
URL: http://opensource.atlassian.com/projects/hibernate/browse/HHH-4577
Project: Hibernate Core
Issue Type: Improvement
Components: caching (L2)
Affects Versions: 3.3.2
Environment: Hibernate 3.3.2GA with JBossCache3.2.1.GA as 2L-cache provider
SQLServer2008
Reporter: Guenther Demetz
Analyzing what our application is mainly doing during flush/commit,
we saw that very often the concerning thread is executing in jbosscache stuff (see 2 sample stacktraces below).
Analyzing the code we discovered that, if doing 1000 inserts for a determinate entity:
-on flush UpdateTimestampsCache.preinvalidate() is called 1000 times for the same space (=table)
-on commit UpdateTimestampsCache.invalidate() is called another 1000 times again for the same space (=table)
It would be much smarter to collect the interested spaces once per flush and once per commit and then execute the (pre)invalidate once per single space. In the case above it would reduce the calls to one single UpdateTimestampsCache.preinvalidate()call on flush and one single UpdateTimestampsCache.invalidate() on commit.
For the commit the enhancement could be following:
ActionQueue.java:
public void afterTransactionCompletion(boolean success) {
int size = executions.size();
final boolean invalidateQueryCache = session.getFactory().getSettings().isQueryCacheEnabled();
Set<Serializable> spaces = invalidateQueryCache ? new HashSet<Serializable>() : null;
try {
for ( int i = 0; i < size; i++ ) {
try {
Executable exec = ( Executable ) executions.get( i );
if ( invalidateQueryCache ) {
Serializable[] nspaces = exec.getPropertySpaces();
for (int j=0; j < nspaces.length; j++)
spaces.add(nspaces[j]);
}
exec.afterTransactionCompletion( success );
}
catch ( CacheException ce ) {
log.error( "could not release a cache lock", ce );
// continue loop
}
catch ( Exception e ) {
throw new AssertionFailure( "Exception releasing cache locks", e );
}
}
}
finally {
session.getFactory().getUpdateTimestampsCache().invalidate(spaces.toArray(new Serializable[]{}));
}
executions.clear();
}
Stack traces:
java.lang.Thread.isInterrupted(Native Method)
java.lang.Thread.interrupted(Thread.java:873)
java.util.concurrent.locks.AbstractQueuedSynchronizer.acquireInterruptibly(AbstractQueuedSynchronizer.java:1134)
java.util.concurrent.locks.ReentrantLock.lockInterruptibly(ReentrantLock.java:312)
java.util.concurrent.LinkedBlockingQueue.put(LinkedBlockingQueue.java:241)
org.jboss.cache.RegionImpl.registerEvictionEvent(RegionImpl.java:249)
org.jboss.cache.RegionImpl.registerEvictionEvent(RegionImpl.java:234)
org.jboss.cache.interceptors.EvictionInterceptor.registerEvictionEventToRegionManager(EvictionInterceptor.java:252)
org.jboss.cache.interceptors.EvictionInterceptor.visitPutKeyValueCommand(EvictionInterceptor.java:109)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.MVCCLockingInterceptor.handlePutKeyValueCommand(MVCCLockingInterceptor.java:102)
org.jboss.cache.interceptors.base.PrePostProcessingCommandInterceptor.visitPutKeyValueCommand(PrePostProcessingCommandInterceptor.java:88)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.base.CommandInterceptor.handleDefault(CommandInterceptor.java:131)
org.jboss.cache.commands.AbstractVisitor.visitPutKeyValueCommand(AbstractVisitor.java:65)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.TxInterceptor.attachGtxAndPassUpChain(TxInterceptor.java:301)
org.jboss.cache.interceptors.TxInterceptor.handleDefault(TxInterceptor.java:283)
org.jboss.cache.commands.AbstractVisitor.visitPutKeyValueCommand(AbstractVisitor.java:65)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.CacheMgmtInterceptor.visitPutKeyValueCommand(CacheMgmtInterceptor.java:119)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.InvocationContextInterceptor.handleAll(InvocationContextInterceptor.java:178)
org.jboss.cache.interceptors.InvocationContextInterceptor.visitPutKeyValueCommand(InvocationContextInterceptor.java:82)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.InterceptorChain.invoke(InterceptorChain.java:287)
org.jboss.cache.invocation.CacheInvocationDelegate.put(CacheInvocationDelegate.java:555)
org.hibernate.cache.jbc2.util.CacheHelper.put(CacheHelper.java:212)
org.hibernate.cache.jbc2.timestamp.TimestampsRegionImpl.put(TimestampsRegionImpl.java:128)
org.hibernate.cache.UpdateTimestampsCache.preinvalidate(UpdateTimestampsCache.java:70)
- locked org.hibernate.cache.UpdateTimestampsCache@1afcaae
org.hibernate.engine.ActionQueue.execute(ActionQueue.java:275)
org.hibernate.engine.ActionQueue.executeActions(ActionQueue.java:263)
org.hibernate.engine.ActionQueue.executeActions(ActionQueue.java:167)
org.hibernate.event.def.AbstractFlushingEventListener.performExecutions(AbstractFlushingEventListener.java:321)
Stack trace:
java.lang.Thread.isInterrupted(Native Method)
java.lang.Thread.interrupted(Thread.java:873)
java.util.concurrent.locks.AbstractQueuedSynchronizer.tryAcquireNanos(AbstractQueuedSynchronizer.java:1158)
java.util.concurrent.locks.ReentrantLock.tryLock(ReentrantLock.java:416)
org.jboss.cache.util.concurrent.locks.AbstractSharedLockContainer.acquireLock(AbstractSharedLockContainer.java:94)
org.jboss.cache.lock.MVCCLockManager.lockAndRecord(MVCCLockManager.java:133)
org.jboss.cache.mvcc.MVCCNodeHelper.acquireLock(MVCCNodeHelper.java:157)
org.jboss.cache.mvcc.MVCCNodeHelper.wrapNodeForWriting(MVCCNodeHelper.java:217)
org.jboss.cache.mvcc.MVCCNodeHelper.wrapNodeForWriting(MVCCNodeHelper.java:186)
org.jboss.cache.interceptors.MVCCLockingInterceptor.handlePutKeyValueCommand(MVCCLockingInterceptor.java:101)
org.jboss.cache.interceptors.base.PrePostProcessingCommandInterceptor.visitPutKeyValueCommand(PrePostProcessingCommandInterceptor.java:88)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.base.CommandInterceptor.handleDefault(CommandInterceptor.java:131)
org.jboss.cache.commands.AbstractVisitor.visitPutKeyValueCommand(AbstractVisitor.java:65)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.TxInterceptor.attachGtxAndPassUpChain(TxInterceptor.java:301)
org.jboss.cache.interceptors.TxInterceptor.handleDefault(TxInterceptor.java:283)
org.jboss.cache.commands.AbstractVisitor.visitPutKeyValueCommand(AbstractVisitor.java:65)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.CacheMgmtInterceptor.visitPutKeyValueCommand(CacheMgmtInterceptor.java:119)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.base.CommandInterceptor.invokeNextInterceptor(CommandInterceptor.java:116)
org.jboss.cache.interceptors.InvocationContextInterceptor.handleAll(InvocationContextInterceptor.java:178)
org.jboss.cache.interceptors.InvocationContextInterceptor.visitPutKeyValueCommand(InvocationContextInterceptor.java:82)
org.jboss.cache.commands.write.PutKeyValueCommand.acceptVisitor(PutKeyValueCommand.java:100)
org.jboss.cache.interceptors.InterceptorChain.invoke(InterceptorChain.java:287)
org.jboss.cache.invocation.CacheInvocationDelegate.put(CacheInvocationDelegate.java:555)
org.hibernate.cache.jbc2.util.CacheHelper.put(CacheHelper.java:212)
org.hibernate.cache.jbc2.timestamp.TimestampsRegionImpl.put(TimestampsRegionImpl.java:128)
org.hibernate.cache.UpdateTimestampsCache.invalidate(UpdateTimestampsCache.java:85)
- locked org.hibernate.cache.UpdateTimestampsCache@1afcaae
org.hibernate.engine.ActionQueue.afterTransactionCompletion(ActionQueue.java:202)
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://opensource.atlassian.com/projects/hibernate/secure/Administrators....
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
12 years, 6 months