/* $Id$ */
/**
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.manifoldcf.crawler.system;
import org.apache.manifoldcf.core.interfaces.*;
import org.apache.manifoldcf.agents.interfaces.*;
import org.apache.manifoldcf.crawler.interfaces.*;
import org.apache.manifoldcf.crawler.system.Logging;
import java.util.*;
import java.lang.reflect.*;
/** This class represents a document cleanup thread. This thread's job is to pull document sets to be cleaned up off of
* a queue, and kill them. It finishes a cleanup set by getting rid of the corresponding rows in the job queue, and from
* carrydown, and from hopcount.
*
* There are very few decisions that this thread needs to make; essentially all the hard thought went into deciding
* what documents to queue in the first place.
*
* The only caveat is that the ingestion API may not be accepting delete requests at the time that this thread wants it
* to be able to accept them. In that case, it's acceptable for the thread to block until the ingestion service is
* functioning again.
*
* Transactions are not much needed for this class; it simply needs to not fail to remove the appropriate jobqueue
* table rows at the end of the delete.
*/
public class DocumentCleanupThread extends Thread
{
public static final String _rcsid = "@(#)$Id$";
// Local data
/** Thread id */
protected final String id;
/** This is a reference to the static main document queue */
protected final DocumentCleanupQueue documentCleanupQueue;
/** Delete thread pool reset manager */
protected final DocCleanupResetManager resetManager;
/** Process ID */
protected final String processID;
/** Constructor.
*@param id is the worker thread id.
*/
public DocumentCleanupThread(String id, DocumentCleanupQueue documentCleanupQueue,
DocCleanupResetManager resetManager, String processID)
throws ManifoldCFException
{
super();
this.id = id;
this.documentCleanupQueue = documentCleanupQueue;
this.resetManager = resetManager;
this.processID = processID;
setName("Document cleanup thread '"+id+"'");
setDaemon(true);
}
public void run()
{
resetManager.registerMe();
try
{
// Create a thread context object.
IThreadContext threadContext = ThreadContextFactory.make();
IIncrementalIngester ingester = IncrementalIngesterFactory.make(threadContext);
IJobManager jobManager = JobManagerFactory.make(threadContext);
IRepositoryConnectionManager connMgr = RepositoryConnectionManagerFactory.make(threadContext);
ITransformationConnectionManager transformationConnectionManager = TransformationConnectionManagerFactory.make(threadContext);
IOutputConnectionManager outputConnectionManager = OutputConnectionManagerFactory.make(threadContext);
IReprioritizationTracker rt = ReprioritizationTrackerFactory.make(threadContext);
IRepositoryConnectorPool repositoryConnectorPool = RepositoryConnectorPoolFactory.make(threadContext);
// Loop
while (true)
{
// Do another try/catch around everything in the loop
try
{
if (Thread.currentThread().isInterrupted())
throw new ManifoldCFException("Interrupted",ManifoldCFException.INTERRUPTED);
// Before we begin, conditionally reset
resetManager.waitForReset(threadContext);
// Once we pull something off the queue, we MUST make sure that
// we update its status, even if there is an exception!!!
// See if there is anything on the queue for me
DocumentCleanupSet dds = documentCleanupQueue.getDocuments();
if (dds == null)
// It's a reset, so recycle
continue;
if (Thread.currentThread().isInterrupted())
throw new ManifoldCFException("Interrupted",ManifoldCFException.INTERRUPTED);
IJobDescription job = dds.getJobDescription();
String connectionName = job.getConnectionName();
IPipelineConnections pipelineConnections = new PipelineConnections(new PipelineSpecificationBasic(job),transformationConnectionManager,outputConnectionManager);
try
{
long currentTime = System.currentTimeMillis();
// Produce a map of connection name->connection object. We will use this to perform a request for multiple connector objects
IRepositoryConnection connection = connMgr.load(connectionName);
// This is where we store the hopcount cleanup data
List<String> arrayDocHashes = new ArrayList<String>();
List<CleanupQueuedDocument> arrayDocsToDelete = new ArrayList<CleanupQueuedDocument>();
List<String[]> arrayRelationshipTypes = new ArrayList<String[]>();
List<Integer> hopcountMethods = new ArrayList<Integer>();
for (int j = 0; j < dds.getCount(); j++)
{
CleanupQueuedDocument dqd = dds.getDocument(j);
DocumentDescription ddd = dqd.getDocumentDescription();
if (job != null && connection != null)
{
// We'll need the legal link types; grab those before we proceed
String[] legalLinkTypes = RepositoryConnectorFactory.getRelationshipTypes(threadContext,connection.getClassName());
if (legalLinkTypes != null)
{
arrayDocHashes.add(ddd.getDocumentIdentifierHash());
arrayDocsToDelete.add(dqd);
arrayRelationshipTypes.add(legalLinkTypes);
hopcountMethods.add(new Integer(job.getHopcountMode()));
}
}
}
// Grab one connection for each connectionName. If we fail, nothing is lost and retries are possible.
IRepositoryConnector connector = repositoryConnectorPool.grab(connection);
try
{
// Iterate over the outputs
boolean[] deleteFromQueue = new boolean[arrayDocHashes.size()];
// Count the number of docs to actually delete. This will be a subset of the documents in the list.
int removeCount = 0;
for (int k = 0; k < arrayDocHashes.size(); k++)
{
if (arrayDocsToDelete.get(k).shouldBeRemovedFromIndex())
{
deleteFromQueue[k] = false;
removeCount++;
}
else
deleteFromQueue[k] = true;
}
// Allocate removal arrays
String[] docClassesToRemove = new String[removeCount];
String[] hashedDocsToRemove = new String[removeCount];
// Now, iterate over the list
removeCount = 0;
for (int k = 0; k < arrayDocHashes.size(); k++)
{
if (arrayDocsToDelete.get(k).shouldBeRemovedFromIndex())
{
docClassesToRemove[removeCount] = connectionName;
hashedDocsToRemove[removeCount] = arrayDocHashes.get(k);
removeCount++;
}
}
OutputRemoveActivity activities = new OutputRemoveActivity(connectionName,connMgr);
// Finally, go ahead and delete the documents from the ingestion system.
try
{
ingester.documentDeleteMultiple(pipelineConnections,docClassesToRemove,hashedDocsToRemove,activities);
// Success! Label all these as needing deletion from queue.
for (int k = 0; k < arrayDocHashes.size(); k++)
{
if (arrayDocsToDelete.get(k).shouldBeRemovedFromIndex())
deleteFromQueue[k] = true;
}
}
catch (ServiceInterruption e)
{
// We don't know which failed, or maybe they all did.
// Go through the list of documents we just tried, and reset them on the queue based on the
// ServiceInterruption parameters. Then we must proceed to delete ONLY the documents that
// were not part of the index deletion attempt.
for (int k = 0; k < arrayDocHashes.size(); k++)
{
CleanupQueuedDocument cqd = arrayDocsToDelete.get(k);
if (cqd.shouldBeRemovedFromIndex())
{
DocumentDescription dd = cqd.getDocumentDescription();
// To recover from a cleanup failure, requeue the document to PURGATORY.
jobManager.resetCleaningDocument(dd,e.getRetryTime());
cqd.setProcessed();
}
}
}
// Successfully deleted some documents from ingestion system. Now, remove them from job queue. This
// must currently happen one document at a time, because the jobs and connectors for each document
// potentially differ.
for (int k = 0; k < arrayDocHashes.size(); k++)
{
if (deleteFromQueue[k])
{
CleanupQueuedDocument dqd = arrayDocsToDelete.get(k);
DocumentDescription ddd = dqd.getDocumentDescription();
Long jobID = ddd.getJobID();
int hopcountMethod = hopcountMethods.get(k).intValue();
String[] legalLinkTypes = arrayRelationshipTypes.get(k);
DocumentDescription[] requeueCandidates = jobManager.markDocumentCleanedUp(jobID,legalLinkTypes,ddd,hopcountMethod);
// Use the common method for doing the requeuing
ManifoldCF.requeueDocumentsDueToCarrydown(jobManager,requeueCandidates,
connector,connection,rt,currentTime);
// Finally, completed expiration of the document.
dqd.setProcessed();
}
}
}
finally
{
// Free up the reserved connector instance
repositoryConnectorPool.release(connection,connector);
}
}
catch (ManifoldCFException e)
{
if (e.getErrorCode() == ManifoldCFException.INTERRUPTED)
break;
if (e.getErrorCode() == ManifoldCFException.DATABASE_CONNECTION_ERROR)
throw e;
// It's ok to abort a job because we can't talk to the search engine during cleanup.
if (jobManager.errorAbort(dds.getJobDescription().getID(),e.getMessage()))
{
// We eat the exception if there was already one recorded.
// An exception occurred in the processing of a set of documents.
// Shut the corresponding job down, with an appropriate error
Logging.threads.error("Exception tossed: "+e.getMessage(),e);
}
}
finally
{
// Insure that the documents that were not deleted get restored to the proper state.
int j = 0;
while (j < dds.getCount())
{
DeleteQueuedDocument dqd = dds.getDocument(j);
if (dqd.wasProcessed() == false)
{
DocumentDescription ddd = dqd.getDocumentDescription();
// Requeue this document!
jobManager.resetCleaningDocument(ddd,0L);
dqd.setProcessed();
}
j++;
}
}
}
catch (ManifoldCFException e)
{
if (e.getErrorCode() == ManifoldCFException.INTERRUPTED)
break;
if (e.getErrorCode() == ManifoldCFException.DATABASE_CONNECTION_ERROR)
{
// Note the failure, which will cause a reset to occur
resetManager.noteEvent();
Logging.threads.error("Document cleanup thread aborting and restarting due to database connection reset: "+e.getMessage(),e);
try
{
// Give the database a chance to catch up/wake up
ManifoldCF.sleep(10000L);
}
catch (InterruptedException se)
{
break;
}
continue;
}
// An exception occurred in the cleanup from another error.
// Log the error (but that's all we can do)
Logging.threads.error("Exception tossed: "+e.getMessage(),e);
}
catch (InterruptedException e)
{
// We're supposed to quit
break;
}
catch (OutOfMemoryError e)
{
System.err.println("agents process ran out of memory - shutting down");
e.printStackTrace(System.err);
System.exit(-200);
}
catch (Throwable e)
{
// A more severe error - but stay alive
Logging.threads.fatal("Error tossed: "+e.getMessage(),e);
}
}
}
catch (Throwable e)
{
// Severe error on initialization
System.err.println("agents process could not start - shutting down");
Logging.threads.fatal("DocumentCleanupThread "+id+" initialization error tossed: "+e.getMessage(),e);
System.exit(-300);
}
}
/** The ingest logger class */
protected static class OutputRemoveActivity implements IOutputRemoveActivity
{
// Connection name
protected final String connectionName;
// Connection manager
protected final IRepositoryConnectionManager connMgr;
/** Constructor */
public OutputRemoveActivity(String connectionName, IRepositoryConnectionManager connMgr)
{
this.connectionName = connectionName;
this.connMgr = connMgr;
}
/** Record time-stamped information about the activity of the output connector.
*@param startTime is either null or the time since the start of epoch in milliseconds (Jan 1, 1970). Every
* activity has an associated time; the startTime field records when the activity began. A null value
* indicates that the start time and the finishing time are the same.
*@param activityType is a string which is fully interpretable only in the context of the connector involved, which is
* used to categorize what kind of activity is being recorded. For example, a web connector might record a
* "fetch document" activity. Cannot be null.
*@param dataSize is the number of bytes of data involved in the activity, or null if not applicable.
*@param entityURI is a (possibly long) string which identifies the object involved in the history record.
* The interpretation of this field will differ from connector to connector. May be null.
*@param resultCode contains a terse description of the result of the activity. The description is limited in
* size to 255 characters, and can be interpreted only in the context of the current connector. May be null.
*@param resultDescription is a (possibly long) human-readable string which adds detail, if required, to the result
* described in the resultCode field. This field is not meant to be queried on. May be null.
*/
public void recordActivity(Long startTime, String activityType, Long dataSize,
String entityURI, String resultCode, String resultDescription)
throws ManifoldCFException
{
connMgr.recordHistory(connectionName,startTime,activityType,dataSize,entityURI,resultCode,
resultDescription,null);
}
}
}