[jira] Commented: (JCR-1572) DbDataStore connection does not always reconnect

["Brennan Spies (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/JCR-1572?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12630649#action_12630649 ] 

Brennan Spies commented on JCR-1572:
------------------------------------

I haven't looked at the patch yet, but I am experiencing a very similar problem using Apache Derby. It is occurring because I am using Artifactory 1.3 (beta) which updates its Maven indexes using Quartz scheduled jobs that rely on Apache Jackrabbit to store/persist the data. The code path of the problem is the same--through the addRecord() method, which automatically calls setAutoReconnect(false) without checking for a valid connection. Forgive me if I am repeating the above discussion, but there are two problems here:

1. Not checking for valid connection before setting autoReconnect(false). This seems to be happening constantly in the Artifactory jobs, since they are long-running and the connections tend to time out. Some logic equivalent to DBCP's testOnBorrow() needs to be implemented.
2. The setAutoReconnect(false)...setAutoReconnect(true) are both inside a try-catch block. What happens here if an exception is thrown between the two calls? Since the ConnectionRecoveryManager instances are held in a pool (the Pool class), the ConnectionRecoveryManager will be returned to the pool with the autoReconnect flag = false. Assuming that execution continues after the original exception, this ConnectionRecoveryManager instance will be likely to throw another exception if a connection timeout occurs. THE SOLUTION: restore the correct flag using setAutoReconnect in the finally block.

These should be very easy fixes to implement.

The exception looks like this:

[9/12/08 11:05:24:697 PDT] 0000004e SystemOut     O 2008-09-12 11:05:24,696 [WARN ] (data.db.DbDataStore :588)  - Can not insert new record
org.apache.jackrabbit.core.data.DataStoreException: Can not insert new record: connection has been closed and autoReconnect == false: connection has been closed and autoReconnect == false
        at org.apache.jackrabbit.core.data.db.DbDataStore.convert(DbDataStore.java:592)
        at org.apache.jackrabbit.core.data.db.DbDataStore.addRecord(DbDataStore.java:295)
        at org.apache.jackrabbit.core.value.BLOBInDataStore.getInstance(BLOBInDataStore.java:120)
        at org.apache.jackrabbit.core.value.InternalValue.getBLOBFileValue(InternalValue.java:644)
        at org.apache.jackrabbit.core.value.InternalValue.create(InternalValue.java:123)
        at org.apache.jackrabbit.core.PropertyImpl.setValue(PropertyImpl.java:609)
        at org.apache.jackrabbit.core.PropertyImpl.setValue(PropertyImpl.java:525)
        at org.apache.jackrabbit.core.NodeImpl.setProperty(NodeImpl.java:2309)
        at org.artifactory.jcr.fs.JcrFile.fillJcrData(JcrFile.java:591)
        at org.artifactory.jcr.fs.JcrFile.setResourceNode(JcrFile.java:550)
        at org.artifactory.jcr.fs.JcrFile.<init>(JcrFile.java:80)
        at org.artifactory.jcr.fs.JcrFile.<init>(JcrFile.java:164)
        at org.artifactory.repo.index.RepoIndexerData.saveIndexFiles(RepoIndexerData.java:103)
        at org.artifactory.repo.index.IndexerManagerImpl.saveIndexFiles(IndexerManagerImpl.java:168)
        at org.artifactory.repo.index.IndexerManagerImpl$$FastClassByCGLIB$$76058b92.invoke(<generated>)
        at net.sf.cglib.proxy.MethodProxy.invoke(MethodProxy.java:149)
        at org.springframework.aop.framework.Cglib2AopProxy$CglibMethodInvocation.invokeJoinpoint(Cglib2AopProxy.java:700)
        at org.springframework.aop.framework.ReflectiveMethodInvocation.proceed(ReflectiveMethodInvocation.java:149)
        at org.springframework.transaction.interceptor.TransactionInterceptor.invoke(TransactionInterceptor.java:106)
        at org.springframework.aop.framework.ReflectiveMethodInvocation.proceed(ReflectiveMethodInvocation.java:171)
        at org.springframework.aop.framework.Cglib2AopProxy$DynamicAdvisedInterceptor.intercept(Cglib2AopProxy.java:635)
        at org.artifactory.repo.index.IndexerManagerImpl$$EnhancerByCGLIB$$7cded8f3.saveIndexFiles(<generated>)
        at org.artifactory.repo.index.IndexerManagerImpl.index(IndexerManagerImpl.java:137)
        at org.artifactory.repo.index.IndexerManagerImpl$$FastClassByCGLIB$$76058b92.invoke(<generated>)
        at net.sf.cglib.proxy.MethodProxy.invoke(MethodProxy.java:149)
        at org.springframework.aop.framework.Cglib2AopProxy$CglibMethodInvocation.invokeJoinpoint(Cglib2AopProxy.java:700)
        at org.springframework.aop.framework.ReflectiveMethodInvocation.proceed(ReflectiveMethodInvocation.java:149)
        at org.springframework.transaction.interceptor.TransactionInterceptor.invoke(TransactionInterceptor.java:106)
        at org.springframework.aop.framework.ReflectiveMethodInvocation.proceed(ReflectiveMethodInvocation.java:171)
        at org.springframework.aop.framework.Cglib2AopProxy$DynamicAdvisedInterceptor.intercept(Cglib2AopProxy.java:635)
        at org.artifactory.repo.index.IndexerManagerImpl$$EnhancerByCGLIB$$7cded8f3.index(<generated>)
        at org.artifactory.repo.index.IndexerJob.onExecute(IndexerJob.java:25)
        at org.artifactory.schedule.ArtifactoryJob.execute(ArtifactoryJob.java:41)
        at org.quartz.core.JobRunShell.run(JobRunShell.java:202)
        at org.quartz.simpl.SimpleThreadPool$WorkerThread.run(SimpleThreadPool.java:525)
Caused by: java.sql.SQLException: connection has been closed and autoReconnect == false
        at org.apache.jackrabbit.core.persistence.bundle.util.ConnectionRecoveryManager.getConnection(ConnectionRecoveryManager.java:159)
        at org.apache.jackrabbit.core.persistence.bundle.util.ConnectionRecoveryManager.executeStmtInternal(ConnectionRecoveryManager.java:287)
        at org.apache.jackrabbit.core.persistence.bundle.util.ConnectionRecoveryManager.executeStmt(ConnectionRecoveryManager.java:257)
        at org.apache.jackrabbit.core.persistence.bundle.util.ConnectionRecoveryManager.executeStmt(ConnectionRecoveryManager.java:237)
        at org.apache.jackrabbit.core.data.db.DbDataStore.addRecord(DbDataStore.java:286)
        ... 33 more



> DbDataStore connection does not always reconnect
> ------------------------------------------------
>
>                 Key: JCR-1572
>                 URL: https://issues.apache.org/jira/browse/JCR-1572
>             Project: Jackrabbit
>          Issue Type: Bug
>          Components: jackrabbit-core
>            Reporter: Moshe Immerman
>            Assignee: Thomas Mueller
>         Attachments: jackrabbit-trunk.patch, jackrabbit1.4.patch
>
>
> If a DbDataStore connection is closed due to an error all subsequent addRecord calls will fail with 'connection has been closed and autoReconnect == false'
>  after getRecord is called and the connection is reconnected addRecord will succeed.
> the connection should be validated before setting autoReconnect = false or on retrieval from the pool.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.