[jboss-user] [JBoss Cache: Core Edition] - BerkeleyDB Deadlock Exception from JBoss Cache Loader: Bdbje

niuxuetao_fiserv do-not-reply at jboss.com
Sat Jun 20 05:23:23 EDT 2009


Hi,

I am using Jboss Pojo cache on top of Jboss cache (3.0.0GA) managing rather big object trees of collections. And I am randomly getting com.sleepycat.je.DeadlockException(s) when using the BdbjeCacheLoader as cache loader. 

There are 10 threads running under a single machine (JSE environment), each does both reading and writing. Neither reading nor writing (as I believe) should take up to 20 seconds as a transactional batch on my hardware, since the objects are not huge (each object is a list of other object that include an array of maps of basic types). 

If I set the lockAcquisitionTimeout to smaller value, it happens more frequently. But even I set the value to a very big one, like 20000, it still happens with less frequency, and it blocks for 20 seconds before this exception is thrown.

Due to these posts, I guess it might have something to do with mvcc. (I am using je-3.3.82), but really do not hope so, since mvcc is one of the points I like jboss cache.

http://forums.oracle.com/forums/thread.jspa?threadID=866063&tstart=60
http://forums.oracle.com/forums/thread.jspa?threadID=616358&tstart=30

Below is the jboss cache configuration and the exception stack trace. Thanks in advance for your help!


  | <?xml version="1.0" encoding="UTF-8"?>
  | <jbosscache xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  | 	xmlns="urn:jboss:jbosscache-core:config:3.0">
  | 
  | 	<locking isolationLevel="READ_COMMITTED"
  | 		lockParentForChildInsertRemove="false" lockAcquisitionTimeout="5000"
  | 		nodeLockingScheme="mvcc" writeSkewCheck="false" concurrencyLevel="500" />
  | 
  | 	<transaction
  | 		transactionManagerLookupClass="org.jboss.cache.transaction.GenericTransactionManagerLookup"
  | 		syncRollbackPhase="false" syncCommitPhase="false" />
  | 
  | 	<jmxStatistics enabled="false" />
  | 
  | 	<startup regionsInactiveOnStartup="true" />
  | 
  | 	<shutdown hookBehavior="DEFAULT" />
  | 
  | 	<listeners asyncPoolSize="1" asyncQueueSize="100000" />
  | 
  | 	<invocationBatching enabled="false" />
  | 
  | 	<eviction wakeUpInterval="1000">
  | 
  | 		<!-- Cache wide default -->
  | 		<region name="/_default_" algorithmClass="org.jboss.cache.eviction.LRUAlgorithm">
  | 
  | 			<property name="maxNodes">10000</property>
  | 			<property name="timeToLiveSeconds">300</property>
  | 		</region>
  | 	</eviction>
  | 
  | 	<loaders passivation="false" shared="true">
  | 		<preload>
  | 			<node fqn="/" />
  | 		</preload>
  | 
  | 		<loader class="org.jboss.cache.loader.bdbje.BdbjeCacheLoader" async="false"
  | 			fetchPersistentState="false" ignoreModifications="false"
  | 			purgeOnStartup="false">
  | 
  | 			<properties>
  | 
  | 				cache.jdbc.table.name=Cached_Events
  | 				cache.jdbc.table.create=false
  | 				cache.jdbc.table.drop=false
  | 				cache.jdbc.table.primarykey=jbosscache_pk
  | 				cache.jdbc.fqn.column=fqn
  | 				cache.jdbc.fqn.type=varchar2(255)
  | 				cache.jdbc.node.column=node
  | 				cache.jdbc.node.type=blob
  | 				cache.jdbc.parent.column=parent
  | 				
  | 				cache.jdbc.sql-concat=concat(1,2)
  | 				
  | 				cache.jdbc.connection.factory=org.jboss.cache.loader.C3p0ConnectionFactory
  | 				c3p0.maxPoolSize=20
  | 				c3p0.checkoutTimeout=5000
  | 				
  | 				location=tmp/jbosscache
  | 			</properties>
  | 
  | 		</loader>
  | 
  | 
  | 	</loaders>
  | 
  | </jbosscache>
  | 


  | com.sleepycat.je.DeadlockException: (JE 3.3.82) Lock expired. Locker 6761395 -1_Detection2_ThreadLocker: waited for lock on database=JBossCache-Cluster LockAddr:23471145 node=1785 type=READ grant=WAIT_NEW timeoutMillis=20000 startTime=1245426236094 endTime=1245426256097
  |     [junit] Owners: [<LockInfo locker="12706060 346_Detection1_Txn" type="WRITE"/>]
  |     [junit] Waiters: [<LockInfo locker="24478649 347_Detection4_Txn" type="WRITE"/>]
  |     [junit] 
  |     [junit] 	at com.sleepycat.je.txn.LockManager.makeTimeoutMsgInternal(LockManager.java:469)
  |     [junit] 	at com.sleepycat.je.txn.SyncedLockManager.makeTimeoutMsg(SyncedLockManager.java:77)
  |     [junit] 	at com.sleepycat.je.txn.LockManager.lock(LockManager.java:278)
  |     [junit] 	at com.sleepycat.je.txn.BasicLocker.lockInternal(BasicLocker.java:167)
  |     [junit] 	at com.sleepycat.je.txn.Locker.lock(Locker.java:360)
  |     [junit] 	at com.sleepycat.je.dbi.CursorImpl.lockLNDeletedAllowed(CursorImpl.java:2516)
  |     [junit] 	at com.sleepycat.je.dbi.CursorImpl.lockLN(CursorImpl.java:2438)
  |     [junit] 	at com.sleepycat.je.dbi.CursorImpl.fetchCurrent(CursorImpl.java:2368)
  |     [junit] 	at com.sleepycat.je.dbi.CursorImpl.getCurrentAlreadyLatched(CursorImpl.java:1427)
  |     [junit] 	at com.sleepycat.je.dbi.CursorImpl.getNextWithKeyChangeStatus(CursorImpl.java:1573)
  |     [junit] 	at com.sleepycat.je.dbi.CursorImpl.getNext(CursorImpl.java:1499)
  |     [junit] 	at com.sleepycat.je.dbi.CursorImpl.getNextNoDup(CursorImpl.java:1688)
  |     [junit] 	at com.sleepycat.je.Cursor.searchInternal(Cursor.java:1936)
  |     [junit] 	at com.sleepycat.je.Cursor.searchAllowPhantoms(Cursor.java:1748)
  |     [junit] 	at com.sleepycat.je.Cursor.search(Cursor.java:1615)
  |     [junit] 	at com.sleepycat.je.Cursor.getSearchKeyRange(Cursor.java:1117)
  |     [junit] 	at org.jboss.cache.loader.bdbje.BdbjeCacheLoader.getChildrenNames(BdbjeCacheLoader.java:391)
  |     [junit] 	at org.jboss.cache.loader.CacheLoaderManager.preload(CacheLoaderManager.java:371)
  | 

View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4238872#4238872

Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4238872



More information about the jboss-user mailing list