Package org.nuxeo.lib.stream.log

Interface LogTailer<M extends Externalizable>

All Superinterfaces:

AutoCloseable

All Known Implementing Classes:

ChronicleCompoundLogTailer, ChronicleLogTailer, KafkaLogTailer

public interface LogTailer<M extends Externalizable>
extends AutoCloseable

Sequential reader for a partition or multiple partitions. A tailer is not thread safe and should not be shared by multiple threads.

Method Summary

Modifier and Type	Method	Description
Collection<LogPartition>	assignments()	Returns the list of Log name, partitions tuples currently assigned to this tailer.
void	close()
boolean	closed()	Returns true if the tailer has been closed.
void	commit()	Commit current positions for all partitions (last message offset returned by read).
LogOffset	commit(LogPartition partition)	Commit current position for the partition.
Codec<M>	getCodec()	Returns the codec used to read the records.
Name	group()	Returns the consumer group.
LogOffset	offsetForTimestamp(LogPartition partition, long timestamp)	Look up the offset for the given partition by timestamp.
LogRecord<M>	read(Duration timeout)	Read a message from assigned partitions within the timeout.
void	reset()	Reset all committed positions for this group, next read will be done from beginning.
void	reset(LogPartition partition)	Reset the committed position for this group on this partition, next read for this partition will be done from the beginning.
void	seek(LogOffset offset)	Set the current position for a single partition.
void	toEnd()	Set the current positions to the end of all partitions.
void	toLastCommitted()	Set the current positions to previously committed positions.
void	toStart()	Set the current positions to the beginning of all partitions.

Method Detail

group

Name group()

Returns the consumer group.

assignments

Collection<LogPartition> assignments()

Returns the list of Log name, partitions tuples currently assigned to this tailer. Assignments can change only if the tailer has been created using LogManager.subscribe(org.nuxeo.lib.stream.log.Name, java.util.Collection<org.nuxeo.lib.stream.log.Name>, org.nuxeo.lib.stream.log.RebalanceListener, org.nuxeo.lib.stream.codec.Codec<M>).

read

LogRecord<M> read(Duration timeout)
           throws InterruptedException

Read a message from assigned partitions within the timeout.

Returns:

null if there is no message in the queue after the timeout.

Throws:

RebalanceException - if a partition rebalancing happen during the read, this is possible only when using LogManager.subscribe(org.nuxeo.lib.stream.log.Name, java.util.Collection<org.nuxeo.lib.stream.log.Name>, org.nuxeo.lib.stream.log.RebalanceListener, org.nuxeo.lib.stream.codec.Codec<M>).

InterruptedException

commit

void commit()

Commit current positions for all partitions (last message offset returned by read).

commit

LogOffset commit(LogPartition partition)

Commit current position for the partition.

Returns:

the committed offset, can return null if there was no previous read done on this partition.

toEnd

void toEnd()

Set the current positions to the end of all partitions.

toStart

void toStart()

Set the current positions to the beginning of all partitions.

toLastCommitted

void toLastCommitted()

Set the current positions to previously committed positions.

seek

void seek(LogOffset offset)

Set the current position for a single partition. Do not change other partitions positions.

Since:

9.3

offsetForTimestamp

LogOffset offsetForTimestamp(LogPartition partition,
                             long timestamp)

Look up the offset for the given partition by timestamp. The position is the earliest offset whose timestamp is greater than or equal to the given timestamp.

The timestamp used depends on the implementation, for Kafka this is the LogAppendTime. Returns null if no record offset is found with an appropriate timestamp.

Since:

10.1

reset

void reset()

Reset all committed positions for this group, next read will be done from beginning.

Since:

9.3

reset

void reset(LogPartition partition)

Reset the committed position for this group on this partition, next read for this partition will be done from the beginning.

Since:

9.3

close

void close()

Specified by:

close in interface AutoCloseable

closed

boolean closed()

Returns true if the tailer has been closed.

getCodec

Codec<M> getCodec()

Returns the codec used to read the records. A null codec is the default legacy encoding.

Since:

10.2