/*
* 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.cassandra.cql3.statements;
import java.nio.ByteBuffer;
import java.util.*;
import java.util.concurrent.TimeUnit;
import com.google.common.collect.Iterables;
import com.google.common.collect.Maps;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.slf4j.helpers.MessageFormatter;
import org.apache.cassandra.schema.TableId;
import org.apache.cassandra.schema.TableMetadata;
import org.apache.cassandra.schema.ColumnMetadata;
import org.apache.cassandra.config.DatabaseDescriptor;
import org.apache.cassandra.cql3.*;
import org.apache.cassandra.db.*;
import org.apache.cassandra.db.partitions.PartitionUpdate;
import org.apache.cassandra.db.rows.RowIterator;
import org.apache.cassandra.exceptions.*;
import org.apache.cassandra.metrics.BatchMetrics;
import org.apache.cassandra.service.*;
import org.apache.cassandra.tracing.Tracing;
import org.apache.cassandra.transport.messages.ResultMessage;
import org.apache.cassandra.utils.FBUtilities;
import org.apache.cassandra.utils.NoSpamLogger;
import org.apache.cassandra.utils.Pair;
import static org.apache.cassandra.cql3.statements.RequestValidations.checkFalse;
/**
* A <code>BATCH</code> statement parsed from a CQL query.
*/
public class BatchStatement implements CQLStatement
{
public enum Type
{
LOGGED, UNLOGGED, COUNTER
}
private final int boundTerms;
public final Type type;
private final List<ModificationStatement> statements;
// Columns modified for each table (keyed by the table ID)
private final Map<TableId, RegularAndStaticColumns> updatedColumns;
// Columns on which there is conditions. Note that if there is any, then the batch can only be on a single partition (and thus table).
private final RegularAndStaticColumns conditionColumns;
private final boolean updatesRegularRows;
private final boolean updatesStaticRow;
private final Attributes attrs;
private final boolean hasConditions;
private static final Logger logger = LoggerFactory.getLogger(BatchStatement.class);
private static final String UNLOGGED_BATCH_WARNING = "Unlogged batch covering {} partitions detected " +
"against table{} {}. You should use a logged batch for " +
"atomicity, or asynchronous writes for performance.";
private static final String LOGGED_BATCH_LOW_GCGS_WARNING = "Executing a LOGGED BATCH on table{} {}, configured with a " +
"gc_grace_seconds of 0. The gc_grace_seconds is used to TTL " +
"batchlog entries, so setting gc_grace_seconds too low on " +
"tables involved in an atomic batch might cause batchlog " +
"entries to expire before being replayed.";
public static final BatchMetrics metrics = new BatchMetrics();
/**
* Creates a new BatchStatement.
*
* @param type type of the batch
* @param statements the list of statements in the batch
* @param attrs additional attributes for statement (CL, timestamp, timeToLive)
*/
public BatchStatement(int boundTerms, Type type, List<ModificationStatement> statements, Attributes attrs)
{
this.boundTerms = boundTerms;
this.type = type;
this.statements = statements;
this.attrs = attrs;
boolean hasConditions = false;
MultiTableColumnsBuilder regularBuilder = new MultiTableColumnsBuilder();
RegularAndStaticColumns.Builder conditionBuilder = RegularAndStaticColumns.builder();
boolean updateRegular = false;
boolean updateStatic = false;
for (ModificationStatement stmt : statements)
{
regularBuilder.addAll(stmt.metadata(), stmt.updatedColumns());
updateRegular |= stmt.updatesRegularRows();
if (stmt.hasConditions())
{
hasConditions = true;
conditionBuilder.addAll(stmt.conditionColumns());
updateStatic |= stmt.updatesStaticRow();
}
}
this.updatedColumns = regularBuilder.build();
this.conditionColumns = conditionBuilder.build();
this.updatesRegularRows = updateRegular;
this.updatesStaticRow = updateStatic;
this.hasConditions = hasConditions;
}
public Iterable<org.apache.cassandra.cql3.functions.Function> getFunctions()
{
List<org.apache.cassandra.cql3.functions.Function> functions = new ArrayList<>();
for (ModificationStatement statement : statements)
statement.addFunctionsTo(functions);
return functions;
}
public int getBoundTerms()
{
return boundTerms;
}
public void checkAccess(ClientState state) throws InvalidRequestException, UnauthorizedException
{
for (ModificationStatement statement : statements)
statement.checkAccess(state);
}
// Validates a prepared batch statement without validating its nested statements.
public void validate() throws InvalidRequestException
{
if (attrs.isTimeToLiveSet())
throw new InvalidRequestException("Global TTL on the BATCH statement is not supported.");
boolean timestampSet = attrs.isTimestampSet();
if (timestampSet)
{
if (hasConditions)
throw new InvalidRequestException("Cannot provide custom timestamp for conditional BATCH");
if (isCounter())
throw new InvalidRequestException("Cannot provide custom timestamp for counter BATCH");
}
boolean hasCounters = false;
boolean hasNonCounters = false;
for (ModificationStatement statement : statements)
{
if (timestampSet && statement.isCounter())
throw new InvalidRequestException("Cannot provide custom timestamp for a BATCH containing counters");
if (timestampSet && statement.isTimestampSet())
throw new InvalidRequestException("Timestamp must be set either on BATCH or individual statements");
if (isCounter() && !statement.isCounter())
throw new InvalidRequestException("Cannot include non-counter statement in a counter batch");
if (isLogged() && statement.isCounter())
throw new InvalidRequestException("Cannot include a counter statement in a logged batch");
if (statement.isCounter())
hasCounters = true;
else
hasNonCounters = true;
}
if (hasCounters && hasNonCounters)
throw new InvalidRequestException("Counter and non-counter mutations cannot exist in the same batch");
if (hasConditions)
{
String ksName = null;
String cfName = null;
for (ModificationStatement stmt : statements)
{
if (ksName != null && (!stmt.keyspace().equals(ksName) || !stmt.columnFamily().equals(cfName)))
throw new InvalidRequestException("Batch with conditions cannot span multiple tables");
ksName = stmt.keyspace();
cfName = stmt.columnFamily();
}
}
}
private boolean isCounter()
{
return type == Type.COUNTER;
}
private boolean isLogged()
{
return type == Type.LOGGED;
}
// The batch itself will be validated in either Parsed#prepare() - for regular CQL3 batches,
// or in QueryProcessor.processBatch() - for native protocol batches.
public void validate(ClientState state) throws InvalidRequestException
{
for (ModificationStatement statement : statements)
statement.validate(state);
}
public List<ModificationStatement> getStatements()
{
return statements;
}
private Collection<? extends IMutation> getMutations(BatchQueryOptions options, boolean local, long now, long queryStartNanoTime)
throws RequestExecutionException, RequestValidationException
{
Set<String> tablesWithZeroGcGs = null;
UpdatesCollector collector = new UpdatesCollector(updatedColumns, updatedRows());
for (int i = 0; i < statements.size(); i++)
{
ModificationStatement statement = statements.get(i);
if (isLogged() && statement.metadata().params.gcGraceSeconds == 0)
{
if (tablesWithZeroGcGs == null)
tablesWithZeroGcGs = new HashSet<>();
tablesWithZeroGcGs.add(statement.metadata.toString());
}
QueryOptions statementOptions = options.forStatement(i);
long timestamp = attrs.getTimestamp(now, statementOptions);
statement.addUpdates(collector, statementOptions, local, timestamp, queryStartNanoTime);
}
if (tablesWithZeroGcGs != null)
{
String suffix = tablesWithZeroGcGs.size() == 1 ? "" : "s";
NoSpamLogger.log(logger, NoSpamLogger.Level.WARN, 1, TimeUnit.MINUTES, LOGGED_BATCH_LOW_GCGS_WARNING,
suffix, tablesWithZeroGcGs);
ClientWarn.instance.warn(MessageFormatter.arrayFormat(LOGGED_BATCH_LOW_GCGS_WARNING, new Object[] { suffix, tablesWithZeroGcGs })
.getMessage());
}
collector.validateIndexedColumns();
return collector.toMutations();
}
private int updatedRows()
{
// Note: it's possible for 2 statements to actually apply to the same row, but that's just an estimation
// for sizing our PartitionUpdate backing array, so it's good enough.
return statements.size();
}
/**
* Checks batch size to ensure threshold is met. If not, a warning is logged.
*
* @param mutations - the batch mutations.
*/
private static void verifyBatchSize(Collection<? extends IMutation> mutations) throws InvalidRequestException
{
// We only warn for batch spanning multiple mutations (#10876)
if (mutations.size() <= 1)
return;
long warnThreshold = DatabaseDescriptor.getBatchSizeWarnThreshold();
long size = IMutation.dataSize(mutations);
if (size > warnThreshold)
{
Set<String> tableNames = new HashSet<>();
for (IMutation mutation : mutations)
{
for (PartitionUpdate update : mutation.getPartitionUpdates())
tableNames.add(update.metadata().toString());
}
long failThreshold = DatabaseDescriptor.getBatchSizeFailThreshold();
String format = "Batch for {} is of size {}, exceeding specified threshold of {} by {}.{}";
if (size > failThreshold)
{
Tracing.trace(format, tableNames, FBUtilities.prettyPrintMemory(size), FBUtilities.prettyPrintMemory(failThreshold),
FBUtilities.prettyPrintMemory(size - failThreshold), " (see batch_size_fail_threshold_in_kb)");
logger.error(format, tableNames, FBUtilities.prettyPrintMemory(size), FBUtilities.prettyPrintMemory(failThreshold),
FBUtilities.prettyPrintMemory(size - failThreshold), " (see batch_size_fail_threshold_in_kb)");
throw new InvalidRequestException("Batch too large");
}
else if (logger.isWarnEnabled())
{
logger.warn(format, tableNames, FBUtilities.prettyPrintMemory(size), FBUtilities.prettyPrintMemory(warnThreshold),
FBUtilities.prettyPrintMemory(size - warnThreshold), "");
}
ClientWarn.instance.warn(MessageFormatter.arrayFormat(format, new Object[] {tableNames, size, warnThreshold, size - warnThreshold, ""}).getMessage());
}
}
private void verifyBatchType(Collection<? extends IMutation> mutations)
{
if (!isLogged() && mutations.size() > 1)
{
Set<DecoratedKey> keySet = new HashSet<>();
Set<String> tableNames = new HashSet<>();
for (IMutation mutation : mutations)
{
for (PartitionUpdate update : mutation.getPartitionUpdates())
{
keySet.add(update.partitionKey());
tableNames.add(update.metadata().toString());
}
}
// CASSANDRA-11529: log only if we have more than a threshold of keys, this was also suggested in the
// original ticket that introduced this warning, CASSANDRA-9282
if (keySet.size() > DatabaseDescriptor.getUnloggedBatchAcrossPartitionsWarnThreshold())
{
NoSpamLogger.log(logger, NoSpamLogger.Level.WARN, 1, TimeUnit.MINUTES, UNLOGGED_BATCH_WARNING,
keySet.size(), tableNames.size() == 1 ? "" : "s", tableNames);
ClientWarn.instance.warn(MessageFormatter.arrayFormat(UNLOGGED_BATCH_WARNING, new Object[]{keySet.size(),
tableNames.size() == 1 ? "" : "s", tableNames}).getMessage());
}
}
}
public ResultMessage execute(QueryState queryState, QueryOptions options, long queryStartNanoTime) throws RequestExecutionException, RequestValidationException
{
return execute(queryState, BatchQueryOptions.withoutPerStatementVariables(options), queryStartNanoTime);
}
public ResultMessage execute(QueryState queryState, BatchQueryOptions options, long queryStartNanoTime) throws RequestExecutionException, RequestValidationException
{
return execute(queryState, options, false, options.getTimestamp(queryState), queryStartNanoTime);
}
private ResultMessage execute(QueryState queryState, BatchQueryOptions options, boolean local, long now, long queryStartNanoTime)
throws RequestExecutionException, RequestValidationException
{
if (options.getConsistency() == null)
throw new InvalidRequestException("Invalid empty consistency level");
if (options.getSerialConsistency() == null)
throw new InvalidRequestException("Invalid empty serial consistency level");
if (hasConditions)
return executeWithConditions(options, queryState, queryStartNanoTime);
executeWithoutConditions(getMutations(options, local, now, queryStartNanoTime), options.getConsistency(), queryStartNanoTime);
return new ResultMessage.Void();
}
private void executeWithoutConditions(Collection<? extends IMutation> mutations, ConsistencyLevel cl, long queryStartNanoTime) throws RequestExecutionException, RequestValidationException
{
if (mutations.isEmpty())
return;
verifyBatchSize(mutations);
verifyBatchType(mutations);
updatePartitionsPerBatchMetrics(mutations.size());
boolean mutateAtomic = (isLogged() && mutations.size() > 1);
StorageProxy.mutateWithTriggers(mutations, cl, mutateAtomic, queryStartNanoTime);
}
private void updatePartitionsPerBatchMetrics(int updatedPartitions)
{
if (isLogged()) {
metrics.partitionsPerLoggedBatch.update(updatedPartitions);
} else if (isCounter()) {
metrics.partitionsPerCounterBatch.update(updatedPartitions);
} else {
metrics.partitionsPerUnloggedBatch.update(updatedPartitions);
}
}
private ResultMessage executeWithConditions(BatchQueryOptions options, QueryState state, long queryStartNanoTime)
throws RequestExecutionException, RequestValidationException
{
Pair<CQL3CasRequest, Set<ColumnMetadata>> p = makeCasRequest(options, state);
CQL3CasRequest casRequest = p.left;
Set<ColumnMetadata> columnsWithConditions = p.right;
String ksName = casRequest.metadata.keyspace;
String tableName = casRequest.metadata.name;
try (RowIterator result = StorageProxy.cas(ksName,
tableName,
casRequest.key,
casRequest,
options.getSerialConsistency(),
options.getConsistency(),
state.getClientState(),
queryStartNanoTime))
{
return new ResultMessage.Rows(ModificationStatement.buildCasResultSet(ksName,
tableName,
result,
columnsWithConditions,
true,
options.forStatement(0)));
}
}
private Pair<CQL3CasRequest,Set<ColumnMetadata>> makeCasRequest(BatchQueryOptions options, QueryState state)
{
long now = state.getTimestamp();
DecoratedKey key = null;
CQL3CasRequest casRequest = null;
Set<ColumnMetadata> columnsWithConditions = new LinkedHashSet<>();
for (int i = 0; i < statements.size(); i++)
{
ModificationStatement statement = statements.get(i);
QueryOptions statementOptions = options.forStatement(i);
long timestamp = attrs.getTimestamp(now, statementOptions);
List<ByteBuffer> pks = statement.buildPartitionKeyNames(statementOptions);
if (statement.getRestrictions().keyIsInRelation())
throw new IllegalArgumentException("Batch with conditions cannot span multiple partitions (you cannot use IN on the partition key)");
if (key == null)
{
key = statement.metadata().partitioner.decorateKey(pks.get(0));
casRequest = new CQL3CasRequest(statement.metadata(), key, true, conditionColumns, updatesRegularRows, updatesStaticRow);
}
else if (!key.getKey().equals(pks.get(0)))
{
throw new InvalidRequestException("Batch with conditions cannot span multiple partitions");
}
checkFalse(statement.getRestrictions().clusteringKeyRestrictionsHasIN(),
"IN on the clustering key columns is not supported with conditional %s",
statement.type.isUpdate()? "updates" : "deletions");
Clustering clustering = Iterables.getOnlyElement(statement.createClustering(statementOptions));
if (statement.hasConditions())
{
statement.addConditions(clustering, casRequest, statementOptions);
// As soon as we have a ifNotExists, we set columnsWithConditions to null so that everything is in the resultSet
if (statement.hasIfNotExistCondition() || statement.hasIfExistCondition())
columnsWithConditions = null;
else if (columnsWithConditions != null)
Iterables.addAll(columnsWithConditions, statement.getColumnsWithConditions());
}
casRequest.addRowUpdate(clustering, statement, statementOptions, timestamp);
}
return Pair.create(casRequest, columnsWithConditions);
}
public ResultMessage executeInternal(QueryState queryState, QueryOptions options) throws RequestValidationException, RequestExecutionException
{
if (hasConditions)
return executeInternalWithConditions(BatchQueryOptions.withoutPerStatementVariables(options), queryState);
executeInternalWithoutCondition(queryState, options, System.nanoTime());
return new ResultMessage.Void();
}
private ResultMessage executeInternalWithoutCondition(QueryState queryState, QueryOptions options, long queryStartNanoTime) throws RequestValidationException, RequestExecutionException
{
for (IMutation mutation : getMutations(BatchQueryOptions.withoutPerStatementVariables(options), true, queryState.getTimestamp(), queryStartNanoTime))
mutation.apply();
return null;
}
private ResultMessage executeInternalWithConditions(BatchQueryOptions options, QueryState state) throws RequestExecutionException, RequestValidationException
{
Pair<CQL3CasRequest, Set<ColumnMetadata>> p = makeCasRequest(options, state);
CQL3CasRequest request = p.left;
Set<ColumnMetadata> columnsWithConditions = p.right;
String ksName = request.metadata.keyspace;
String tableName = request.metadata.name;
try (RowIterator result = ModificationStatement.casInternal(request, state))
{
return new ResultMessage.Rows(ModificationStatement.buildCasResultSet(ksName, tableName, result, columnsWithConditions, true, options.forStatement(0)));
}
}
public String toString()
{
return String.format("BatchStatement(type=%s, statements=%s)", type, statements);
}
public static class Parsed extends CFStatement
{
private final Type type;
private final Attributes.Raw attrs;
private final List<ModificationStatement.Parsed> parsedStatements;
public Parsed(Type type, Attributes.Raw attrs, List<ModificationStatement.Parsed> parsedStatements)
{
super(null);
this.type = type;
this.attrs = attrs;
this.parsedStatements = parsedStatements;
}
@Override
public void prepareKeyspace(ClientState state) throws InvalidRequestException
{
for (ModificationStatement.Parsed statement : parsedStatements)
statement.prepareKeyspace(state);
}
public ParsedStatement.Prepared prepare() throws InvalidRequestException
{
VariableSpecifications boundNames = getBoundVariables();
String firstKS = null;
String firstCF = null;
boolean haveMultipleCFs = false;
List<ModificationStatement> statements = new ArrayList<>(parsedStatements.size());
for (ModificationStatement.Parsed parsed : parsedStatements)
{
if (firstKS == null)
{
firstKS = parsed.keyspace();
firstCF = parsed.columnFamily();
}
else if (!haveMultipleCFs)
{
haveMultipleCFs = !firstKS.equals(parsed.keyspace()) || !firstCF.equals(parsed.columnFamily());
}
statements.add(parsed.prepare(boundNames));
}
Attributes prepAttrs = attrs.prepare("[batch]", "[batch]");
prepAttrs.collectMarkerSpecification(boundNames);
BatchStatement batchStatement = new BatchStatement(boundNames.size(), type, statements, prepAttrs);
batchStatement.validate();
// Use the TableMetadata of the first statement for partition key bind indexes. If the statements affect
// multiple tables, we won't send partition key bind indexes.
short[] partitionKeyBindIndexes = (haveMultipleCFs || batchStatement.statements.isEmpty())? null
: boundNames.getPartitionKeyBindIndexes(batchStatement.statements.get(0).metadata());
return new ParsedStatement.Prepared(batchStatement, boundNames, partitionKeyBindIndexes);
}
}
private static class MultiTableColumnsBuilder
{
private final Map<TableId, RegularAndStaticColumns.Builder> perTableBuilders = new HashMap<>();
public void addAll(TableMetadata table, RegularAndStaticColumns columns)
{
RegularAndStaticColumns.Builder builder = perTableBuilders.get(table.id);
if (builder == null)
{
builder = RegularAndStaticColumns.builder();
perTableBuilders.put(table.id, builder);
}
builder.addAll(columns);
}
public Map<TableId, RegularAndStaticColumns> build()
{
Map<TableId, RegularAndStaticColumns> m = Maps.newHashMapWithExpectedSize(perTableBuilders.size());
for (Map.Entry<TableId, RegularAndStaticColumns.Builder> p : perTableBuilders.entrySet())
m.put(p.getKey(), p.getValue().build());
return m;
}
}
}