Search in sources :

Example 1 with PartitionConsumerResult

use of io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumerResult in project cdap by caskdata.

the class PartitionConsumerTest method testNumRetries.

@Test
public void testNumRetries() throws Exception {
    final PartitionedFileSet dataset = dsFrameworkUtil.getInstance(pfsInstance);
    final TransactionAware txAwareDataset = (TransactionAware) dataset;
    final int numRetries = 1;
    ConsumerConfiguration configuration = ConsumerConfiguration.builder().setMaxRetries(numRetries).build();
    final PartitionConsumer partitionConsumer = new ConcurrentPartitionConsumer(dataset, new InMemoryStatePersistor(), configuration);
    final PartitionKey partitionKey = generateUniqueKey();
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            dataset.getPartitionOutput(partitionKey).addPartition();
        }
    });
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            // from the working set
            for (int i = 0; i < numRetries + 1; i++) {
                List<PartitionDetail> partitionDetails = partitionConsumer.consumePartitions(1).getPartitions();
                Assert.assertEquals(1, partitionDetails.size());
                Assert.assertEquals(partitionKey, partitionDetails.get(0).getPartitionKey());
                // aborting the processing of the partition
                partitionConsumer.onFinish(partitionDetails, false);
            }
            // after the 2nd abort, the partition is discarded entirely, and so no partitions are available for consuming
            PartitionConsumerResult result = partitionConsumer.consumePartitions(1);
            Assert.assertEquals(0, result.getPartitions().size());
            Assert.assertEquals(1, result.getFailedPartitions().size());
            Assert.assertEquals(partitionKey, result.getFailedPartitions().get(0).getPartitionKey());
        }
    });
}
Also used : ConcurrentPartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.ConcurrentPartitionConsumer) PartitionConsumerResult(io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumerResult) PartitionedFileSet(io.cdap.cdap.api.dataset.lib.PartitionedFileSet) TransactionExecutor(org.apache.tephra.TransactionExecutor) TransactionAware(org.apache.tephra.TransactionAware) ConsumerConfiguration(io.cdap.cdap.api.dataset.lib.partitioned.ConsumerConfiguration) PartitionKey(io.cdap.cdap.api.dataset.lib.PartitionKey) ArrayList(java.util.ArrayList) ImmutableList(com.google.common.collect.ImmutableList) List(java.util.List) PartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumer) ConcurrentPartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.ConcurrentPartitionConsumer) Test(org.junit.Test)

Example 2 with PartitionConsumerResult

use of io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumerResult in project cdap by caskdata.

the class PartitionConsumerTest method testPartitionPutback.

@Test
public void testPartitionPutback() throws Exception {
    final PartitionedFileSet dataset = dsFrameworkUtil.getInstance(pfsInstance);
    final TransactionAware txAwareDataset = (TransactionAware) dataset;
    final Set<PartitionKey> partitionKeys = new HashSet<>();
    for (int i = 0; i < 10; i++) {
        partitionKeys.add(generateUniqueKey());
    }
    final PartitionConsumer partitionConsumer = new ConcurrentPartitionConsumer(dataset, new InMemoryStatePersistor(), ConsumerConfiguration.builder().setMaxRetries(1).build());
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            for (PartitionKey partitionKey : partitionKeys) {
                dataset.getPartitionOutput(partitionKey).addPartition();
            }
        }
    });
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            // consume all the partitions
            List<? extends Partition> consumedPartitions = partitionConsumer.consumePartitions().getPartitions();
            Assert.assertEquals(partitionKeys, toKeys(consumedPartitions));
            // consuming the partitions again, without adding any new partitions returns an empty iterator
            Assert.assertTrue(partitionConsumer.consumePartitions().getPartitions().isEmpty());
            // and testing that they are still available for processing, and that there are no failed partitions
            for (int i = 0; i < 5; i++) {
                partitionConsumer.untake(consumedPartitions);
                PartitionConsumerResult result = partitionConsumer.consumePartitions();
                consumedPartitions = result.getPartitions();
                Assert.assertEquals(partitionKeys, toKeys(consumedPartitions));
                Assert.assertEquals(0, result.getFailedPartitions().size());
            }
            // consuming the partitions again, without adding any new partitions returns an empty iterator
            Assert.assertTrue(partitionConsumer.consumePartitions().getPartitions().isEmpty());
            // test functionality to put back a partial subset of the retrieved the partitions
            Partition firstConsumedPartition = consumedPartitions.get(0);
            // test the untakeWithKeys method
            partitionConsumer.untakeWithKeys(ImmutableList.of(firstConsumedPartition.getPartitionKey()));
            consumedPartitions = partitionConsumer.consumePartitions().getPartitions();
            Assert.assertEquals(1, consumedPartitions.size());
            Assert.assertEquals(firstConsumedPartition, consumedPartitions.get(0));
        }
    });
}
Also used : ConcurrentPartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.ConcurrentPartitionConsumer) ConsumablePartition(io.cdap.cdap.api.dataset.lib.partitioned.ConsumablePartition) Partition(io.cdap.cdap.api.dataset.lib.Partition) PartitionConsumerResult(io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumerResult) PartitionedFileSet(io.cdap.cdap.api.dataset.lib.PartitionedFileSet) TransactionExecutor(org.apache.tephra.TransactionExecutor) TransactionAware(org.apache.tephra.TransactionAware) PartitionKey(io.cdap.cdap.api.dataset.lib.PartitionKey) ArrayList(java.util.ArrayList) ImmutableList(com.google.common.collect.ImmutableList) List(java.util.List) PartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumer) ConcurrentPartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.ConcurrentPartitionConsumer) HashSet(java.util.HashSet) Test(org.junit.Test)

Example 3 with PartitionConsumerResult

use of io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumerResult in project cdap by caskdata.

the class PartitionConsumerTest method testDroppedPartitions.

@Test
public void testDroppedPartitions() throws Exception {
    // Tests the case of a partition in the partition consumer working set being dropped from the Partitioned
    // FileSet (See CDAP-6215)
    final PartitionedFileSet dataset = dsFrameworkUtil.getInstance(pfsInstance);
    final TransactionAware txAwareDataset = (TransactionAware) dataset;
    ConsumerConfiguration configuration = ConsumerConfiguration.builder().setMaxWorkingSetSize(1).setMaxRetries(2).build();
    final PartitionConsumer partitionConsumer = new ConcurrentPartitionConsumer(dataset, new InMemoryStatePersistor(), configuration);
    final PartitionKey partitionKey1 = generateUniqueKey();
    final PartitionKey partitionKey2 = generateUniqueKey();
    // Note: These two partitions are added in separate transactions, so that the first can exist in the working set
    // without the second. Partitions in the same transaction can not be split up (due to their index being the same)
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            dataset.getPartitionOutput(partitionKey1).addPartition();
        }
    });
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            dataset.getPartitionOutput(partitionKey2).addPartition();
        }
    });
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            // consuming and aborting the partition numRetries times plus one (for the first attempt) makes it get removed
            // from the working set
            List<PartitionDetail> partitionDetails = partitionConsumer.consumePartitions(1).getPartitions();
            Assert.assertEquals(1, partitionDetails.size());
            Assert.assertEquals(partitionKey1, partitionDetails.get(0).getPartitionKey());
            // aborting the processing of the partition, to put it back in the working set
            partitionConsumer.onFinish(partitionDetails, false);
        }
    });
    // dropping partitionKey1 from the dataset makes it no longer available for consuming
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            dataset.dropPartition(partitionKey1);
        }
    });
    dsFrameworkUtil.newInMemoryTransactionExecutor(txAwareDataset).execute(new TransactionExecutor.Subroutine() {

        @Override
        public void apply() throws Exception {
            // first call to consume will drop the partition from the working set, and return nothing, since it was
            // the only partition in the working set
            PartitionConsumerResult result = partitionConsumer.consumePartitions(1);
            Assert.assertEquals(0, result.getPartitions().size());
            Assert.assertEquals(0, result.getFailedPartitions().size());
            // following calls to consumePartitions will repopulate the working set and return additional partition(s)
            result = partitionConsumer.consumePartitions(1);
            Assert.assertEquals(1, result.getPartitions().size());
            Assert.assertEquals(partitionKey2, result.getPartitions().get(0).getPartitionKey());
        }
    });
}
Also used : ConcurrentPartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.ConcurrentPartitionConsumer) PartitionConsumerResult(io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumerResult) PartitionedFileSet(io.cdap.cdap.api.dataset.lib.PartitionedFileSet) TransactionExecutor(org.apache.tephra.TransactionExecutor) TransactionAware(org.apache.tephra.TransactionAware) ConsumerConfiguration(io.cdap.cdap.api.dataset.lib.partitioned.ConsumerConfiguration) PartitionKey(io.cdap.cdap.api.dataset.lib.PartitionKey) ArrayList(java.util.ArrayList) ImmutableList(com.google.common.collect.ImmutableList) List(java.util.List) PartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumer) ConcurrentPartitionConsumer(io.cdap.cdap.api.dataset.lib.partitioned.ConcurrentPartitionConsumer) Test(org.junit.Test)

Aggregations

ImmutableList (com.google.common.collect.ImmutableList)3 PartitionKey (io.cdap.cdap.api.dataset.lib.PartitionKey)3 PartitionedFileSet (io.cdap.cdap.api.dataset.lib.PartitionedFileSet)3 ConcurrentPartitionConsumer (io.cdap.cdap.api.dataset.lib.partitioned.ConcurrentPartitionConsumer)3 PartitionConsumer (io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumer)3 PartitionConsumerResult (io.cdap.cdap.api.dataset.lib.partitioned.PartitionConsumerResult)3 ArrayList (java.util.ArrayList)3 List (java.util.List)3 TransactionAware (org.apache.tephra.TransactionAware)3 TransactionExecutor (org.apache.tephra.TransactionExecutor)3 Test (org.junit.Test)3 ConsumerConfiguration (io.cdap.cdap.api.dataset.lib.partitioned.ConsumerConfiguration)2 Partition (io.cdap.cdap.api.dataset.lib.Partition)1 ConsumablePartition (io.cdap.cdap.api.dataset.lib.partitioned.ConsumablePartition)1 HashSet (java.util.HashSet)1