private String runResolveCommand(List<String> args) {
int loopCount = 0;
if (args.size() == 0) {
return null;
}
StringBuffer allOutput = new StringBuffer();
int numProcessed = 0;
if (maxArgs < MIN_ALLOWABLE_ARGS) {
LOG.warn(
"Invalid value "
+ Integer.toString(maxArgs)
+ " for "
+ SCRIPT_ARG_COUNT_KEY
+ "; must be >= "
+ Integer.toString(MIN_ALLOWABLE_ARGS));
return null;
}
while (numProcessed != args.size()) {
int start = maxArgs * loopCount;
List<String> cmdList = new ArrayList<String>();
cmdList.add(scriptName);
for (numProcessed = start;
numProcessed < (start + maxArgs) && numProcessed < args.size();
numProcessed++) {
cmdList.add(args.get(numProcessed));
}
File dir = null;
String userDir;
if ((userDir = System.getProperty("user.dir")) != null) {
dir = new File(userDir);
}
ShellCommandExecutor s = new ShellCommandExecutor(cmdList.toArray(new String[0]), dir);
try {
s.execute();
allOutput.append(s.getOutput() + " ");
} catch (Exception e) {
LOG.warn(StringUtils.stringifyException(e));
return null;
}
loopCount++;
}
return allOutput.toString();
}
/**
* list subfiles
*
* @param hdfsPath !null path probably exists
* @return !null list of enclused file names - emptu if file of not exists
*/
@Override
public String[] ls(final String hdfsPath) {
if (isRunningAsUser()) {
return super.ls(hdfsPath);
}
final FileSystem fs = getDFS();
try {
FileStatus[] statuses = fs.listStatus(new Path(hdfsPath));
if (statuses == null) return new String[0];
List<String> holder = new ArrayList<String>();
for (int i = 0; i < statuses.length; i++) {
FileStatus statuse = statuses[i];
String s = statuse.getPath().getName();
holder.add(s);
}
String[] ret = new String[holder.size()];
holder.toArray(ret);
return ret;
} catch (IOException e) {
throw new RuntimeException(e);
}
}
/**
* Search for pages matching a query, eliminating excessive hits with matching values for a named
* field. Hits after the first <code>maxHitsPerDup</code> are removed from results. The remaining
* hits have {@link Hit#moreFromDupExcluded()} set.
*
* <p>If maxHitsPerDup is zero then all hits are returned.
*
* @param query query
* @param numHits number of requested hits
* @param maxHitsPerDup the maximum hits returned with matching values, or zero
* @param dedupField field name to check for duplicates
* @param sortField Field to sort on (or null if no sorting).
* @param reverse True if we are to reverse sort by <code>sortField</code>.
* @return Hits the matching hits
* @throws IOException
*/
public Hits search(
Query query,
int numHits,
int maxHitsPerDup,
String dedupField,
String sortField,
boolean reverse)
throws IOException {
if (maxHitsPerDup <= 0) // disable dup checking
return search(query, numHits, dedupField, sortField, reverse);
final float rawHitsFactor = this.conf.getFloat("searcher.hostgrouping.rawhits.factor", 2.0f);
int numHitsRaw = (int) (numHits * rawHitsFactor);
if (LOG.isInfoEnabled()) {
LOG.info("searching for " + numHitsRaw + " raw hits");
}
Hits hits = searchBean.search(query, numHitsRaw, dedupField, sortField, reverse);
final long total = hits.getTotal();
final Map<String, DupHits> dupToHits = new HashMap<String, DupHits>();
final List<Hit> resultList = new ArrayList<Hit>();
final Set<Hit> seen = new HashSet<Hit>();
final List<String> excludedValues = new ArrayList<String>();
boolean totalIsExact = true;
for (int rawHitNum = 0; rawHitNum < hits.getTotal(); rawHitNum++) {
// get the next raw hit
if (rawHitNum >= hits.getLength()) {
// optimize query by prohibiting more matches on some excluded values
final Query optQuery = (Query) query.clone();
for (int i = 0; i < excludedValues.size(); i++) {
if (i == MAX_PROHIBITED_TERMS) break;
optQuery.addProhibitedTerm(excludedValues.get(i), dedupField);
}
numHitsRaw = (int) (numHitsRaw * rawHitsFactor);
if (LOG.isInfoEnabled()) {
LOG.info("re-searching for " + numHitsRaw + " raw hits, query: " + optQuery);
}
hits = searchBean.search(optQuery, numHitsRaw, dedupField, sortField, reverse);
if (LOG.isInfoEnabled()) {
LOG.info("found " + hits.getTotal() + " raw hits");
}
rawHitNum = -1;
continue;
}
final Hit hit = hits.getHit(rawHitNum);
if (seen.contains(hit)) continue;
seen.add(hit);
// get dup hits for its value
final String value = hit.getDedupValue();
DupHits dupHits = dupToHits.get(value);
if (dupHits == null) dupToHits.put(value, dupHits = new DupHits());
// does this hit exceed maxHitsPerDup?
if (dupHits.size() == maxHitsPerDup) { // yes -- ignore the hit
if (!dupHits.maxSizeExceeded) {
// mark prior hits with moreFromDupExcluded
for (int i = 0; i < dupHits.size(); i++) {
dupHits.get(i).setMoreFromDupExcluded(true);
}
dupHits.maxSizeExceeded = true;
excludedValues.add(value); // exclude dup
}
totalIsExact = false;
} else { // no -- collect the hit
resultList.add(hit);
dupHits.add(hit);
// are we done?
// we need to find one more than asked for, so that we can tell if
// there are more hits to be shown
if (resultList.size() > numHits) break;
}
}
final Hits results = new Hits(total, resultList.toArray(new Hit[resultList.size()]));
results.setTotalIsExact(totalIsExact);
return results;
}
public int run(String[] args) throws Exception {
// printUsage();
/*
* SETUP
*/
Configuration argConf = getConf();
Hashtable<String, String> confArg = new Hashtable<String, String>();
setup(confArg, argConf);
Date currentTime = new Date();
Date endDate = new Date(new Long(confArg.get("timestamp_stop")));
Boolean full_run = confArg.get("intermediate").matches("(?i).*true.*");
Boolean quick_add = confArg.get("quick_add").matches("(?i).*true.*");
logger.info("Running GeStore");
// ZooKeeper setup
Configuration config = HBaseConfiguration.create();
zkWatcher = new ZooKeeperWatcher(config, "Testing", new HBaseAdmin(config));
zkInstance =
new ZooKeeper(
ZKConfig.getZKQuorumServersString(config),
config.getInt("zookeeper.session.timeout", -1),
zkWatcher);
if (!confArg.get("task_id").isEmpty()) {
confArg.put("temp_path", confArg.get("temp_path") + confArg.get("task_id"));
}
String lockRequest = confArg.get("file_id");
if (!confArg.get("run_id").isEmpty())
lockRequest = lockRequest + "_" + confArg.get("run_id") + "_";
if (!confArg.get("task_id").isEmpty())
lockRequest = lockRequest + "_" + confArg.get("task_id") + "_";
// Get type of movement
toFrom type_move = checkArgs(confArg);
if (type_move == toFrom.LOCAL2REMOTE && !confArg.get("format").equals("unknown")) {
List<String> arguments = new ArrayList<String>();
arguments.add("-Dinput=" + confArg.get("local_path"));
arguments.add("-Dtable=" + confArg.get("file_id"));
arguments.add("-Dtimestamp=" + confArg.get("timestamp_stop"));
arguments.add("-Dtype=" + confArg.get("format"));
arguments.add("-Dtarget_dir=" + confArg.get("base_path") + "_" + confArg.get("file_id"));
arguments.add("-Dtemp_hdfs_path=" + confArg.get("temp_path"));
arguments.add("-Drun_id=" + confArg.get("run_id"));
if (!confArg.get("run_id").isEmpty()) arguments.add("-Drun_id=" + confArg.get("run_id"));
if (!confArg.get("task_id").isEmpty()) arguments.add("-Dtask_id=" + confArg.get("task_id"));
if (quick_add) arguments.add("-Dquick_add=" + confArg.get("quick_add"));
String lockName = lock(lockRequest);
String[] argumentString = arguments.toArray(new String[arguments.size()]);
adddb.main(argumentString);
unlock(lockName);
System.exit(0);
}
// Database registration
dbutil db_util = new dbutil(config);
db_util.register_database(confArg.get("db_name_files"), true);
db_util.register_database(confArg.get("db_name_runs"), true);
db_util.register_database(confArg.get("db_name_updates"), true);
FileSystem hdfs = FileSystem.get(config);
FileSystem localFS = FileSystem.getLocal(config);
// Get source type
confArg.put("source", getSource(db_util, confArg.get("db_name_files"), confArg.get("file_id")));
confArg.put(
"database", isDatabase(db_util, confArg.get("db_name_files"), confArg.get("file_id")));
if (!confArg.get("source").equals("local")
&& type_move == toFrom.REMOTE2LOCAL
&& !confArg.get("timestamp_stop").equals(Integer.toString(Integer.MAX_VALUE))) {
confArg.put("timestamp_stop", Long.toString(latestVersion(confArg, db_util)));
}
/*
* Get previous timestamp
*/
Get run_id_get = new Get(confArg.get("run_id").getBytes());
Result run_get = db_util.doGet(confArg.get("db_name_runs"), run_id_get);
KeyValue run_file_prev =
run_get.getColumnLatest(
"d".getBytes(), (confArg.get("file_id") + "_db_timestamp").getBytes());
String last_timestamp = new String("0");
if (null != run_file_prev && !confArg.get("source").equals("local")) {
long last_timestamp_real = run_file_prev.getTimestamp();
Long current_timestamp = new Long(confArg.get("timestamp_real"));
if ((current_timestamp - last_timestamp_real) > 36000) {
last_timestamp = new String(run_file_prev.getValue());
Integer lastTimestamp = new Integer(last_timestamp);
lastTimestamp += 1;
last_timestamp = lastTimestamp.toString();
logger.info("Last timestamp: " + last_timestamp + " End data: " + endDate);
Date last_run = new Date(run_file_prev.getTimestamp());
if (last_run.before(endDate) && !full_run) {
confArg.put("timestamp_start", last_timestamp);
}
}
}
Integer tse = new Integer(confArg.get("timestamp_stop"));
Integer tss = new Integer(confArg.get("timestamp_start"));
if (tss > tse) {
logger.info("No new version of requested file.");
return 0;
}
/*
* Generate file
*/
String lockName = lock(lockRequest);
Get file_id_get = new Get(confArg.get("file_id").getBytes());
Result file_get = db_util.doGet(confArg.get("db_name_files"), file_id_get);
if (!file_get.isEmpty()) {
boolean found =
hasFile(
db_util,
hdfs,
confArg.get("db_name_files"),
confArg.get("file_id"),
getFullPath(confArg));
if (confArg.get("source").equals("fullfile")) {
found = false;
}
String filenames_put =
getFileNames(
db_util, confArg.get("db_name_files"), confArg.get("file_id"), getFullPath(confArg));
// Filename not found in file database
if (!found && type_move == toFrom.REMOTE2LOCAL) {
if (!confArg.get("source").equals("local")) {
// Generate intermediate file
if (getFile(hdfs, confArg, db_util) == null) {
unlock(lockName);
return 1;
}
// Put generated file into file database
if (!confArg.get("format").equals("fullfile")) {
putFileEntry(
db_util,
hdfs,
confArg.get("db_name_files"),
confArg.get("file_id"),
confArg.get("full_file_name"),
confArg.get("source"));
}
} else {
logger.warn("Remote file not found, and cannot be generated! File: " + confArg);
unlock(lockName);
return 1;
}
}
} else {
if (type_move == toFrom.REMOTE2LOCAL) {
logger.warn("Remote file not found, and cannot be generated.");
unlock(lockName);
return 1;
}
}
/*
* Copy file
* Update tables
*/
if (type_move == toFrom.LOCAL2REMOTE) {
if (!confArg.get("format").equals("fullfile")) {
putFileEntry(
db_util,
hdfs,
confArg.get("db_name_files"),
confArg.get("file_id"),
getFullPath(confArg),
confArg.get("source"));
}
putRunEntry(
db_util,
confArg.get("db_name_runs"),
confArg.get("run_id"),
confArg.get("file_id"),
confArg.get("type"),
confArg.get("timestamp_real"),
confArg.get("timestamp_stop"),
getFullPath(confArg),
confArg.get("delimiter"));
hdfs.copyFromLocalFile(new Path(confArg.get("local_path")), new Path(getFullPath(confArg)));
} else if (type_move == toFrom.REMOTE2LOCAL) {
FileStatus[] files = hdfs.globStatus(new Path(getFullPath(confArg) + "*"));
putRunEntry(
db_util,
confArg.get("db_name_runs"),
confArg.get("run_id"),
confArg.get("file_id"),
confArg.get("type"),
confArg.get("timestamp_real"),
confArg.get("timestamp_stop"),
getFullPath(confArg),
confArg.get("delimiter"));
unlock(lockName);
for (FileStatus file : files) {
Path cur_file = file.getPath();
Path cur_local_path =
new Path(new String(confArg.get("local_path") + confArg.get("file_id")));
String suffix = getSuffix(getFileName(confArg), cur_file.getName());
if (suffix.length() > 0) {
cur_local_path = cur_local_path.suffix(new String("." + suffix));
}
if (confArg.get("copy").equals("true")) {
String crc = hdfs.getFileChecksum(cur_file).toString();
if (checksumLocalTest(cur_local_path, crc)) {
continue;
} else {
hdfs.copyToLocalFile(cur_file, cur_local_path);
writeChecksum(cur_local_path, crc);
}
} else {
System.out.println(cur_local_path + "\t" + cur_file);
}
}
}
unlock(lockName);
return 0;
}
