org.apache.mesos

Class Log

java.lang.Object

org.apache.mesos.Log

public class Log
extends Object

Provides access to a distributed append only log. The log can be read from using a Log.Reader and written to using a Log.Writer.

Both the Reader and Writer will require a quorum which defines the ratio of active Mesos Masters that need to be available for a successful read or write. The quorum will be satisfied when the number of active Masters is greater than the given number:

Quorum > (Number of Masters)/2
 

If a read or write is executed the operation will wait until their is quorum to succeed.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Log.Entry Represents an opaque data entry in the Log with a Log.Position.
static class	Log.OperationFailedException An exception that gets thrown when an error occurs while performing a read or write operation.
static class	Log.Position An opaque identifier of a log entry's position within the log.
static class	Log.Reader Provides read access to the Log.
static class	Log.Writer Provides write access to the Log.
static class	Log.WriterFailedException An exception that gets thrown when a writer no longer has the ability to perform operations (e.g., because it was superseded by another writer).

Constructor Summary

Constructors

Constructor and Description
Log(int quorum, String path, Set<String> pids) Creates a new replicated log that assumes the specified quorum size, is backed by a file at the specified path, and coordiantes with other replicas via the set of process PIDs.
Log(int quorum, String path, String servers, long timeout, TimeUnit unit, String znode) Creates a new replicated log that assumes the specified quorum size, is backed by a file at the specified path, and coordiantes with other replicas associated with the specified ZooKeeper servers, timeout, and znode (or Zookeeper name space).
Log(int quorum, String path, String servers, long timeout, TimeUnit unit, String znode, String scheme, byte[] credentials) Creates a new replicated log that assumes the specified quorum size, is backed by a file at the specified path, and coordiantes with other replicas associated with the specified ZooKeeper servers, timeout, and znode (or Zookeeper name space).

Method Summary

Methods

Modifier and Type	Method and Description
protected void	finalize()
protected void	initialize(int quorum, String path, Set<String> pids)
protected void	initialize(int quorum, String path, String servers, long timeout, TimeUnit unit, String znode)
protected void	initialize(int quorum, String path, String servers, long timeout, TimeUnit unit, String znode, String scheme, byte[] credentials)
Log.Position	position(byte[] identity) Returns a position based off of the bytes recovered from Position.identity().

Methods inherited from class java.lang.Object

clone, equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Log

public Log(int quorum,
   String path,
   Set<String> pids)

Creates a new replicated log that assumes the specified quorum size, is backed by a file at the specified path, and coordiantes with other replicas via the set of process PIDs.

Parameters:

quorum - The quorum size.

path - Path to the file backing this log.

pids - PIDs of the replicas to coordinate with.

Log

public Log(int quorum,
   String path,
   String servers,
   long timeout,
   TimeUnit unit,
   String znode)

Creates a new replicated log that assumes the specified quorum size, is backed by a file at the specified path, and coordiantes with other replicas associated with the specified ZooKeeper servers, timeout, and znode (or Zookeeper name space).

Parameters:

quorum - The quorum size.

path - Path to the file backing this log.

servers - List of ZooKeeper servers (e.g., 'ip1:port1,ip2:port2').

timeout - Max number of time units to wait before a TimeoutException.

unit - Type of time units used for the timeout, e.g. seconds, minutes, etc.

znode - Path to znode where "state" should be rooted.

Log

public Log(int quorum,
   String path,
   String servers,
   long timeout,
   TimeUnit unit,
   String znode,
   String scheme,
   byte[] credentials)

Creates a new replicated log that assumes the specified quorum size, is backed by a file at the specified path, and coordiantes with other replicas associated with the specified ZooKeeper servers, timeout, and znode (or Zookeeper name space).

Parameters:

quorum - The quorum size.

path - Path to the file backing this log.

servers - Zookeper servers/connection string.

timeout - Max number of time units to wait before a TimeoutException.

unit - Type of time units used for the timeout, e.g. seconds, minutes, etc.

znode - The Zookeeper name space.

scheme - Authentication scheme (e.g., "digest").

credentials - Authentication credentials (e.g., "user:pass").

Method Detail

position

public Log.Position position(byte[] identity)

Returns a position based off of the bytes recovered from Position.identity().

Parameters:

identity - Identity, in bytes, of the position.

Returns:

The position.

initialize

protected void initialize(int quorum,
              String path,
              Set<String> pids)

initialize

protected void initialize(int quorum,
              String path,
              String servers,
              long timeout,
              TimeUnit unit,
              String znode)

initialize

protected void initialize(int quorum,
              String path,
              String servers,
              long timeout,
              TimeUnit unit,
              String znode,
              String scheme,
              byte[] credentials)

finalize

protected void finalize()

Overrides:

finalize in class Object