/**
* Calls the Unix sort command with the options <code>$filesNames -o
* $outputfile -T WaybackSettings#WAYBACK_AGGREGATOR_TEMP_DIR.
*
* Sets the LC_ALL environment variable before making the call.
*
* @param files The files to merge and sort
* @param outputFile The resulting sorted file
* @param additionalArgs A list af extra arguments, which (if different from
* null) are added to the sort call.<p> Note: If any
* of the args contain a whitespace the call will
* fail.
*/
private void processFiles(File[] files, File outputFile, List<String> additionalArgs) {
if (files.length == 0) {
// Empty file list will cause sort to wait for further input,
// and the call will therefore never return
return;
}
Process p = null;
try {
List<String> inputFileList = new LinkedList<String>();
for (int i = 0; i < files.length; i++) {
if (files[i].exists() && files[i].isFile()) {
inputFileList.add(files[i].getCanonicalPath());
} else {
log.warn(
"File "
+ files[i]
+ " doesn't exist or isn't a regular file, "
+ "dropping from list of files to "
+ "sort and merge");
}
}
List<String> cmd = new LinkedList<String>();
// Prepare to run the unix sort command, see sort manual page for
// details
cmd.add("sort");
cmd.addAll(inputFileList);
cmd.add("-o");
cmd.add(outputFile.getCanonicalPath());
cmd.add("-T");
cmd.add(Settings.get(WaybackSettings.WAYBACK_AGGREGATOR_TEMP_DIR));
if (additionalArgs != null && !additionalArgs.isEmpty()) {
for (String argument : additionalArgs) {
ArgumentNotValid.checkTrue(
argument.indexOf(' ') == -1,
"The argument '" + argument + "' contains spaces, this isn't allowed ");
}
cmd.addAll(additionalArgs);
}
ProcessBuilder pb = new ProcessBuilder(cmd);
// Reset all locale definitions
pb.environment().put("LC_ALL", "C");
// Run the command in the user.dir directory
pb.directory(new File(System.getProperty("user.dir")));
p = pb.start();
p.waitFor();
if (p.exitValue() != 0) {
log.error("Failed to sort index files, sort exited with " + "return code " + p.exitValue());
}
} catch (Exception e) {
log.error("Failed to aggregate indexes ", e);
}
}
public class IndexerTestCase {
private String oldClient = System.getProperty(CommonSettings.ARC_REPOSITORY_CLIENT);
private String oldFileDir = System.getProperty("settings.common.arcrepositoryClient.fileDir");
protected static File tempdir = new File(Settings.get(WaybackSettings.WAYBACK_INDEX_TEMPDIR));
ReloadSettings rs = new ReloadSettings();
@Before
public void setUp() {
rs.setUp();
System.setProperty(WaybackSettings.HIBERNATE_HBM2DDL_AUTO, "create-drop");
HibernateUtil.getSession().getSessionFactory().close();
FileUtils.removeRecursively(TestInfo.WORKING_DIR);
TestFileUtils.copyDirectoryNonCVS(TestInfo.ORIGINALS_DIR, TestInfo.WORKING_DIR);
System.setProperty(
CommonSettings.ARC_REPOSITORY_CLIENT,
"dk.netarkivet.common.distribute.arcrepository.LocalArcRepositoryClient");
System.setProperty(
"settings.common.arcrepositoryClient.fileDir", TestInfo.FILE_DIR.getAbsolutePath());
System.setProperty(
CommonSettings.REMOTE_FILE_CLASS, "dk.netarkivet.common.distribute.TestRemoteFile");
assertTrue(
ArcRepositoryClientFactory.getPreservationInstance() instanceof LocalArcRepositoryClient);
}
@After
public void tearDown() {
HibernateUtil.getSession().getSessionFactory().close();
FileUtils.removeRecursively(TestInfo.WORKING_DIR);
if (oldClient != null) {
System.setProperty(CommonSettings.ARC_REPOSITORY_CLIENT, oldClient);
} else {
System.setProperty(CommonSettings.ARC_REPOSITORY_CLIENT, "");
}
if (oldFileDir != null) {
System.setProperty("settings.common.arcrepositoryClient.fileDir", oldFileDir);
} else {
System.setProperty("settings.common.arcrepositoryClient.fileDir", "");
}
rs.tearDown();
}
}
/**
* Package private constructor for common initialisation.
*
* @param harvestID the id of the harvestdefinition
* @param cfg the configuration to base the Job on
* @param orderXMLdoc
* @param channel the channel on which the job will be submitted.
* @param forceMaxObjectsPerDomain the maximum number of objects harvested from a domain,
* overrides individual configuration settings. -1 means no limit
* @param forceMaxBytesPerDomain The maximum number of objects harvested from a domain, or -1 for
* no limit.
* @param forceMaxJobRunningTime The max time in seconds given to the harvester for this job
* @param harvestNum the run number of the harvest definition
* @throws ArgumentNotValid if cfg or priority is null or harvestID is invalid, or if any limit <
* -1
*/
public Job(
Long harvestID,
DomainConfiguration cfg,
HeritrixTemplate orderXMLdoc,
HarvestChannel channel,
long forceMaxObjectsPerDomain,
long forceMaxBytesPerDomain,
long forceMaxJobRunningTime,
int harvestNum)
throws ArgumentNotValid {
ArgumentNotValid.checkNotNull(cfg, "cfg");
ArgumentNotValid.checkNotNull(harvestID, "harvestID");
ArgumentNotValid.checkNotNegative(harvestID, "harvestID");
ArgumentNotValid.checkNotNull(channel, "channel");
if (forceMaxObjectsPerDomain < -1) {
String msg = "forceMaxObjectsPerDomain must be either -1 or positive";
log.debug(msg);
throw new ArgumentNotValid(msg);
}
if (forceMaxBytesPerDomain < -1) {
String msg = "forceMaxBytesPerDomain must be either -1 or positive";
log.debug(msg);
throw new ArgumentNotValid(msg);
}
if (forceMaxBytesPerDomain == 0L) {
log.warn(
"forceMaxBytesPerDomain should probably not be 0.Means 0 bytes downloaded per domain");
}
if (forceMaxObjectsPerDomain == 0L) {
log.warn(
"forceMaxObjectsPerDomain should probably not be 0.Means 0 objects downloaded per domain");
}
// setup initial members
domainConfigurationMap = new HashMap<>();
origHarvestDefinitionID = harvestID;
orderXMLname = cfg.getOrderXmlName();
this.orderXMLdoc = orderXMLdoc;
setHarvestChannel(channel);
long maxObjects = NumberUtils.minInf(forceMaxObjectsPerDomain, cfg.getMaxObjects());
setMaxObjectsPerDomain(maxObjects);
configurationSetsObjectLimit = (maxObjects != forceMaxObjectsPerDomain);
long maxBytes = NumberUtils.minInf(forceMaxBytesPerDomain, cfg.getMaxBytes());
setMaxBytesPerDomain(maxBytes);
configurationSetsByteLimit = (maxBytes != forceMaxBytesPerDomain);
long expectation =
cfg.getExpectedNumberOfObjects(forceMaxObjectsPerDomain, forceMaxBytesPerDomain);
maxCountObjects = expectation;
minCountObjects = expectation;
this.harvestNum = harvestNum;
addConfiguration(cfg);
setMaxJobRunningTime(forceMaxJobRunningTime);
setArchiveFormatInTemplate(Settings.get(HarvesterSettings.HERITRIX_ARCHIVE_FORMAT));
setAttributes(cfg.getAttributesAndTypes());
orderXMLdoc.enableOrDisableDeduplication(
Settings.getBoolean(HarvesterSettings.DEDUPLICATION_ENABLED));
status = JobStatus.NEW;
}
/**
* Batchjob that extracts lines from a crawl log matching a regular expression The batch job should
* be restricted to run on metadata files for a specific job only, using the {@link
* #processOnlyFilesMatching(String)} construct.
*/
@SuppressWarnings({"serial"})
public class CrawlLogLinesMatchingRegexp extends ArchiveBatchJob {
/** The logger. */
// private final Log log = LogFactory.getLog(getClass().getName());
private static final Logger log = LoggerFactory.getLogger(CrawlLogLinesMatchingRegexp.class);
/** Metadata URL for crawl logs. */
private static final String SETUP_URL_FORMAT =
String.format(
"metadata://%s/crawl/logs/crawl.log", Settings.get(CommonSettings.ORGANIZATION));
/** The regular expression to match in the crawl.log line. */
private final String regexp;
/**
* Initialise the batch job.
*
* @param regexp The regexp to match in the crawl.log lines.
*/
public CrawlLogLinesMatchingRegexp(String regexp) {
ArgumentNotValid.checkNotNullOrEmpty(regexp, "regexp");
this.regexp = regexp;
/** One week in milliseconds. */
batchJobTimeout = 7 * Constants.ONE_DAY_IN_MILLIES;
}
/**
* Does nothing, no initialisation is needed.
*
* @param os Not used.
*/
@Override
public void initialize(OutputStream os) {}
@Override
public ArchiveBatchFilter getFilter() {
return new ArchiveBatchFilter("OnlyCrawlLog") {
public boolean accept(ArchiveRecordBase record) {
String URL = record.getHeader().getUrl();
if (URL == null) {
return false;
} else {
return URL.startsWith(SETUP_URL_FORMAT);
}
}
};
}
/**
* Process a record on crawl log concerning the given domain to result.
*
* @param record The record to process.
* @param os The output stream for the result.
* @throws ArgumentNotValid on null parameters
* @throws IOFailure on trouble processing the record.
*/
@Override
public void processRecord(ArchiveRecordBase record, OutputStream os) {
ArgumentNotValid.checkNotNull(record, "ArchiveRecordBase record");
ArgumentNotValid.checkNotNull(os, "OutputStream os");
BufferedReader arcreader = new BufferedReader(new InputStreamReader(record.getInputStream()));
try {
for (String line = arcreader.readLine(); line != null; line = arcreader.readLine()) {
if (line.matches(regexp)) {
os.write(line.getBytes("UTF-8"));
os.write('\n');
}
}
} catch (IOException e) {
throw new IOFailure("Unable to process (w)arc record", e);
} finally {
try {
arcreader.close();
} catch (IOException e) {
log.warn("unable to close arcreader probably", e);
}
}
}
/**
* Does nothing, no finishing is needed.
*
* @param os Not used.
*/
@Override
public void finish(OutputStream os) {}
@Override
public String toString() {
return getClass().getName()
+ ", with arguments: Regexp = "
+ regexp
+ ", Filter = "
+ getFilter();
}
}
