private void binMutations(
MutationSet mutationsToProcess,
Map<String, TabletServerMutations<Mutation>> binnedMutations) {
try {
Set<Entry<String, List<Mutation>>> es = mutationsToProcess.getMutations().entrySet();
for (Entry<String, List<Mutation>> entry : es) {
TabletLocator locator = getLocator(entry.getKey());
String table = entry.getKey();
List<Mutation> tableMutations = entry.getValue();
if (tableMutations != null) {
ArrayList<Mutation> tableFailures = new ArrayList<Mutation>();
locator.binMutations(credentials, tableMutations, binnedMutations, tableFailures);
if (tableFailures.size() > 0) {
failedMutations.add(table, tableFailures);
if (tableFailures.size() == tableMutations.size())
if (!Tables.exists(instance, entry.getKey()))
throw new TableDeletedException(entry.getKey());
else if (Tables.getTableState(instance, table) == TableState.OFFLINE)
throw new TableOfflineException(instance, entry.getKey());
}
}
}
return;
} catch (AccumuloServerException ase) {
updateServerErrors(ase.getServer(), ase);
} catch (AccumuloException ae) {
// assume an IOError communicating with !METADATA tablet
failedMutations.add(mutationsToProcess);
} catch (AccumuloSecurityException e) {
updateAuthorizationFailures(
Collections.singletonMap(
new KeyExtent(new Text(MetadataTable.ID), null, null),
SecurityErrorCode.valueOf(e.getSecurityErrorCode().name())));
} catch (TableDeletedException e) {
updateUnknownErrors(e.getMessage(), e);
} catch (TableOfflineException e) {
updateUnknownErrors(e.getMessage(), e);
} catch (TableNotFoundException e) {
updateUnknownErrors(e.getMessage(), e);
}
// an error ocurred
binnedMutations.clear();
}
private void updateAuthorizationFailures(
Map<KeyExtent, SecurityErrorCode> authorizationFailures) {
if (authorizationFailures.size() > 0) {
// was a table deleted?
HashSet<String> tableIds = new HashSet<String>();
for (KeyExtent ke : authorizationFailures.keySet()) tableIds.add(ke.getTableId().toString());
Tables.clearCache(instance);
for (String tableId : tableIds)
if (!Tables.exists(instance, tableId)) throw new TableDeletedException(tableId);
synchronized (this) {
somethingFailed = true;
mergeAuthorizationFailures(this.authorizationFailures, authorizationFailures);
this.notifyAll();
}
}
}
/**
* Gets the splits of the tables that have been set on the job by reading the metadata table for
* the specified ranges.
*
* @return the splits from the tables based on the ranges.
* @throws java.io.IOException if a table set on the job doesn't exist or an error occurs
* initializing the tablet locator
*/
@Override
public InputSplit[] getSplits(JobConf job, int numSplits) throws IOException {
Level logLevel = getLogLevel(job);
log.setLevel(logLevel);
validateOptions(job);
Random random = new Random();
LinkedList<InputSplit> splits = new LinkedList<InputSplit>();
Map<String, InputTableConfig> tableConfigs = getInputTableConfigs(job);
for (Map.Entry<String, InputTableConfig> tableConfigEntry : tableConfigs.entrySet()) {
String tableName = tableConfigEntry.getKey();
InputTableConfig tableConfig = tableConfigEntry.getValue();
Instance instance = getInstance(job);
String tableId;
// resolve table name to id once, and use id from this point forward
if (DeprecationUtil.isMockInstance(instance)) {
tableId = "";
} else {
try {
tableId = Tables.getTableId(instance, tableName);
} catch (TableNotFoundException e) {
throw new IOException(e);
}
}
Authorizations auths = getScanAuthorizations(job);
String principal = getPrincipal(job);
AuthenticationToken token = getAuthenticationToken(job);
boolean batchScan = InputConfigurator.isBatchScan(CLASS, job);
boolean supportBatchScan =
!(tableConfig.isOfflineScan()
|| tableConfig.shouldUseIsolatedScanners()
|| tableConfig.shouldUseLocalIterators());
if (batchScan && !supportBatchScan)
throw new IllegalArgumentException(
"BatchScanner optimization not available for offline scan, isolated, or local iterators");
boolean autoAdjust = tableConfig.shouldAutoAdjustRanges();
if (batchScan && !autoAdjust)
throw new IllegalArgumentException(
"AutoAdjustRanges must be enabled when using BatchScanner optimization");
List<Range> ranges =
autoAdjust ? Range.mergeOverlapping(tableConfig.getRanges()) : tableConfig.getRanges();
if (ranges.isEmpty()) {
ranges = new ArrayList<Range>(1);
ranges.add(new Range());
}
// get the metadata information for these ranges
Map<String, Map<KeyExtent, List<Range>>> binnedRanges =
new HashMap<String, Map<KeyExtent, List<Range>>>();
TabletLocator tl;
try {
if (tableConfig.isOfflineScan()) {
binnedRanges = binOfflineTable(job, tableId, ranges);
while (binnedRanges == null) {
// Some tablets were still online, try again
// sleep randomly between 100 and 200 ms
sleepUninterruptibly(100 + random.nextInt(100), TimeUnit.MILLISECONDS);
binnedRanges = binOfflineTable(job, tableId, ranges);
}
} else {
tl = InputConfigurator.getTabletLocator(CLASS, job, tableId);
// its possible that the cache could contain complete, but old information about a tables
// tablets... so clear it
tl.invalidateCache();
ClientContext context =
new ClientContext(
getInstance(job),
new Credentials(getPrincipal(job), getAuthenticationToken(job)),
getClientConfiguration(job));
while (!tl.binRanges(context, ranges, binnedRanges).isEmpty()) {
if (!DeprecationUtil.isMockInstance(instance)) {
if (!Tables.exists(instance, tableId)) throw new TableDeletedException(tableId);
if (Tables.getTableState(instance, tableId) == TableState.OFFLINE)
throw new TableOfflineException(instance, tableId);
}
binnedRanges.clear();
log.warn("Unable to locate bins for specified ranges. Retrying.");
// sleep randomly between 100 and 200 ms
sleepUninterruptibly(100 + random.nextInt(100), TimeUnit.MILLISECONDS);
tl.invalidateCache();
}
}
} catch (Exception e) {
throw new IOException(e);
}
HashMap<Range, ArrayList<String>> splitsToAdd = null;
if (!autoAdjust) splitsToAdd = new HashMap<Range, ArrayList<String>>();
HashMap<String, String> hostNameCache = new HashMap<String, String>();
for (Map.Entry<String, Map<KeyExtent, List<Range>>> tserverBin : binnedRanges.entrySet()) {
String ip = tserverBin.getKey().split(":", 2)[0];
String location = hostNameCache.get(ip);
if (location == null) {
InetAddress inetAddress = InetAddress.getByName(ip);
location = inetAddress.getCanonicalHostName();
hostNameCache.put(ip, location);
}
for (Map.Entry<KeyExtent, List<Range>> extentRanges : tserverBin.getValue().entrySet()) {
Range ke = extentRanges.getKey().toDataRange();
if (batchScan) {
// group ranges by tablet to be read by a BatchScanner
ArrayList<Range> clippedRanges = new ArrayList<Range>();
for (Range r : extentRanges.getValue()) clippedRanges.add(ke.clip(r));
BatchInputSplit split =
new BatchInputSplit(tableName, tableId, clippedRanges, new String[] {location});
SplitUtils.updateSplit(split, instance, tableConfig, principal, token, auths, logLevel);
splits.add(split);
} else {
// not grouping by tablet
for (Range r : extentRanges.getValue()) {
if (autoAdjust) {
// divide ranges into smaller ranges, based on the tablets
RangeInputSplit split =
new RangeInputSplit(tableName, tableId, ke.clip(r), new String[] {location});
SplitUtils.updateSplit(
split, instance, tableConfig, principal, token, auths, logLevel);
split.setOffline(tableConfig.isOfflineScan());
split.setIsolatedScan(tableConfig.shouldUseIsolatedScanners());
split.setUsesLocalIterators(tableConfig.shouldUseLocalIterators());
splits.add(split);
} else {
// don't divide ranges
ArrayList<String> locations = splitsToAdd.get(r);
if (locations == null) locations = new ArrayList<String>(1);
locations.add(location);
splitsToAdd.put(r, locations);
}
}
}
}
}
if (!autoAdjust)
for (Map.Entry<Range, ArrayList<String>> entry : splitsToAdd.entrySet()) {
RangeInputSplit split =
new RangeInputSplit(
tableName, tableId, entry.getKey(), entry.getValue().toArray(new String[0]));
SplitUtils.updateSplit(split, instance, tableConfig, principal, token, auths, logLevel);
split.setOffline(tableConfig.isOfflineScan());
split.setIsolatedScan(tableConfig.shouldUseIsolatedScanners());
split.setUsesLocalIterators(tableConfig.shouldUseLocalIterators());
splits.add(split);
}
}
return splits.toArray(new InputSplit[splits.size()]);
}
public static List<KeyValue> scan(ClientContext context, ScanState scanState, int timeOut)
throws ScanTimedOutException, AccumuloException, AccumuloSecurityException,
TableNotFoundException {
TabletLocation loc = null;
Instance instance = context.getInstance();
long startTime = System.currentTimeMillis();
String lastError = null;
String error = null;
int tooManyFilesCount = 0;
long sleepMillis = 100;
List<KeyValue> results = null;
Span span = Trace.start("scan");
try {
while (results == null && !scanState.finished) {
if (Thread.currentThread().isInterrupted()) {
throw new AccumuloException("Thread interrupted");
}
if ((System.currentTimeMillis() - startTime) / 1000.0 > timeOut)
throw new ScanTimedOutException();
while (loc == null) {
long currentTime = System.currentTimeMillis();
if ((currentTime - startTime) / 1000.0 > timeOut) throw new ScanTimedOutException();
Span locateSpan = Trace.start("scan:locateTablet");
try {
loc =
TabletLocator.getLocator(context, scanState.tableId)
.locateTablet(context, scanState.startRow, scanState.skipStartRow, false);
if (loc == null) {
if (!Tables.exists(instance, scanState.tableId))
throw new TableDeletedException(scanState.tableId);
else if (Tables.getTableState(instance, scanState.tableId) == TableState.OFFLINE)
throw new TableOfflineException(instance, scanState.tableId);
error =
"Failed to locate tablet for table : "
+ scanState.tableId
+ " row : "
+ scanState.startRow;
if (!error.equals(lastError)) log.debug("{}", error);
else if (log.isTraceEnabled()) log.trace("{}", error);
lastError = error;
sleepMillis = pause(sleepMillis);
} else {
// when a tablet splits we do want to continue scanning the low child
// of the split if we are already passed it
Range dataRange = loc.tablet_extent.toDataRange();
if (scanState.range.getStartKey() != null
&& dataRange.afterEndKey(scanState.range.getStartKey())) {
// go to the next tablet
scanState.startRow = loc.tablet_extent.getEndRow();
scanState.skipStartRow = true;
loc = null;
} else if (scanState.range.getEndKey() != null
&& dataRange.beforeStartKey(scanState.range.getEndKey())) {
// should not happen
throw new RuntimeException(
"Unexpected tablet, extent : "
+ loc.tablet_extent
+ " range : "
+ scanState.range
+ " startRow : "
+ scanState.startRow);
}
}
} catch (AccumuloServerException e) {
log.debug("Scan failed, server side exception : {}", e.getMessage());
throw e;
} catch (AccumuloException e) {
error = "exception from tablet loc " + e.getMessage();
if (!error.equals(lastError)) log.debug("{}", error);
else if (log.isTraceEnabled()) log.trace("{}", error);
lastError = error;
sleepMillis = pause(sleepMillis);
} finally {
locateSpan.stop();
}
}
Span scanLocation = Trace.start("scan:location");
scanLocation.data("tserver", loc.tablet_location);
try {
results = scan(loc, scanState, context);
} catch (AccumuloSecurityException e) {
Tables.clearCache(instance);
if (!Tables.exists(instance, scanState.tableId))
throw new TableDeletedException(scanState.tableId);
e.setTableInfo(Tables.getPrintableTableInfoFromId(instance, scanState.tableId));
throw e;
} catch (TApplicationException tae) {
throw new AccumuloServerException(loc.tablet_location, tae);
} catch (TSampleNotPresentException tsnpe) {
String message =
"Table "
+ Tables.getPrintableTableInfoFromId(instance, scanState.tableId)
+ " does not have sampling configured or built";
throw new SampleNotPresentException(message, tsnpe);
} catch (NotServingTabletException e) {
error = "Scan failed, not serving tablet " + loc;
if (!error.equals(lastError)) log.debug("{}", error);
else if (log.isTraceEnabled()) log.trace("{}", error);
lastError = error;
TabletLocator.getLocator(context, scanState.tableId).invalidateCache(loc.tablet_extent);
loc = null;
// no need to try the current scan id somewhere else
scanState.scanID = null;
if (scanState.isolated) throw new IsolationException();
sleepMillis = pause(sleepMillis);
} catch (NoSuchScanIDException e) {
error = "Scan failed, no such scan id " + scanState.scanID + " " + loc;
if (!error.equals(lastError)) log.debug("{}", error);
else if (log.isTraceEnabled()) log.trace("{}", error);
lastError = error;
if (scanState.isolated) throw new IsolationException();
scanState.scanID = null;
} catch (TooManyFilesException e) {
error = "Tablet has too many files " + loc + " retrying...";
if (!error.equals(lastError)) {
log.debug("{}", error);
tooManyFilesCount = 0;
} else {
tooManyFilesCount++;
if (tooManyFilesCount == 300) log.warn("{}", error);
else if (log.isTraceEnabled()) log.trace("{}", error);
}
lastError = error;
// not sure what state the scan session on the server side is
// in after this occurs, so lets be cautious and start a new
// scan session
scanState.scanID = null;
if (scanState.isolated) throw new IsolationException();
sleepMillis = pause(sleepMillis);
} catch (TException e) {
TabletLocator.getLocator(context, scanState.tableId)
.invalidateCache(context.getInstance(), loc.tablet_location);
error =
"Scan failed, thrift error "
+ e.getClass().getName()
+ " "
+ e.getMessage()
+ " "
+ loc;
if (!error.equals(lastError)) log.debug("{}", error);
else if (log.isTraceEnabled()) log.trace("{}", error);
lastError = error;
loc = null;
// do not want to continue using the same scan id, if a timeout occurred could cause a
// batch to be skipped
// because a thread on the server side may still be processing the timed out continue scan
scanState.scanID = null;
if (scanState.isolated) throw new IsolationException();
sleepMillis = pause(sleepMillis);
} finally {
scanLocation.stop();
}
}
if (results != null && results.size() == 0 && scanState.finished) {
results = null;
}
return results;
} catch (InterruptedException ex) {
throw new AccumuloException(ex);
} finally {
span.stop();
}
}
