org.jgroups.protocols
Class FD

java.lang.Object
  org.jgroups.stack.Protocol
      org.jgroups.protocols.FD

Direct Known Subclasses:

FD_ICMP, FD_PING

public class FD
extends Protocol

Failure detection based on simple heartbeat protocol. Regularly polls members for liveness. Multicasts SUSPECT messages when a member is not reachable. The simple algorithms works as follows: the membership is known and ordered. Each HB protocol periodically sends an 'are-you-alive' message to its *neighbor*. A neighbor is the next in rank in the membership list, which is recomputed upon a view change. When a response hasn't been received for n milliseconds and m tries, the corresponding member is suspected (and eventually excluded if faulty).

FD starts when it detects (in a view change notification) that there are at least 2 members in the group. It stops running when the membership drops below 2.

When a message is received from the monitored neighbor member, it causes the pinger thread to 'skip' sending the next are-you-alive message. Thus, traffic is reduced.

Author:

Bela Ban

Nested Class Summary

protected class	FD.Broadcaster Task that periodically broadcasts a list of suspected members to the group.
protected class	FD.BroadcastTask
static class	FD.FdHeader
protected class	FD.Monitor Task which periodically checks of the last_ack from ping_dest exceeded timeout and - if yes - broadcasts a SUSPECT message

Field Summary

protected FD.Broadcaster	bcast_task Transmits SUSPECT message until view change or UNSUSPECT is received
protected long	last_ack
protected Address	local_addr
protected java.util.concurrent.locks.Lock	lock
protected int	max_tries
protected java.util.List<Address>	members
protected java.util.concurrent.Future<?>	monitor_future
protected int	num_heartbeats
protected int	num_suspect_events
protected java.util.concurrent.atomic.AtomicInteger	num_tries
protected Address	ping_dest
protected java.util.List<Address>	pingable_mbrs Members from which we select ping_dest.
protected BoundedList<Address>	suspect_history
protected long	timeout
protected TimeScheduler	timer

Fields inherited from class org.jgroups.stack.Protocol

down_prot, ergonomics, id, log, name, stack, stats, up_prot

Constructor Summary

FD()

Method Summary

protected void	computePingDest(Address remove) Computes pingable_mbrs (based on the current membership and the suspected members) and ping_dest
java.lang.Object	down(Event evt) An event is to be sent down the stack.
int	getCurrentNumTries()
java.lang.String	getLocalAddress()
int	getMaxTries()
java.lang.String	getMembers()
int	getNumberOfHeartbeatsSent()
int	getNumSuspectEventsGenerated()
java.lang.String	getPingableMembers()
java.lang.String	getPingDest()
protected Address	getPingDest(java.util.List<Address> mbrs)
long	getTimeout()
void	init() Called after instance has been created (null constructor) and before protocol is started.
boolean	isMonitorRunning()
java.lang.String	printSuspectHistory()
void	resetStats()
protected void	sendHeartbeatResponse(Address dest)
void	setMaxTries(int max_tries)
void	setTimeout(long timeout)
void	startFailureDetection()
protected void	startMonitor() Requires lock to held by caller
void	stop() This method is called on a Channel.disconnect().
void	stopFailureDetection()
protected void	stopMonitor() Requires lock to be held by caller
protected void	unsuspect(Address mbr)
java.lang.Object	up(Event evt) An event was received from the layer below.
protected void	updateTimestamp(Address sender)

Methods inherited from class org.jgroups.stack.Protocol

destroy, dumpStats, enableStats, getConfigurableObjects, getDownProtocol, getDownServices, getId, getIdsAbove, getLevel, getName, getProtocolStack, getSocketFactory, getThreadFactory, getTransport, getUpProtocol, getUpServices, getValue, isErgonomics, printStats, providedDownServices, providedUpServices, requiredDownServices, requiredUpServices, resetStatistics, setDownProtocol, setErgonomics, setId, setLevel, setProtocolStack, setSocketFactory, setUpProtocol, setValue, setValues, start, statsEnabled

Methods inherited from class java.lang.Object

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

Field Detail

timeout

protected long timeout

max_tries

protected int max_tries

num_heartbeats

protected int num_heartbeats

num_suspect_events

protected int num_suspect_events

suspect_history

protected final BoundedList<Address> suspect_history

local_addr

protected Address local_addr

last_ack

protected volatile long last_ack

num_tries

protected final java.util.concurrent.atomic.AtomicInteger num_tries

lock

protected final java.util.concurrent.locks.Lock lock

ping_dest

protected volatile Address ping_dest

members

protected final java.util.List<Address> members

pingable_mbrs

protected final java.util.List<Address> pingable_mbrs

Members from which we select ping_dest. Copy of members minus the suspected members

timer

protected TimeScheduler timer

monitor_future

protected java.util.concurrent.Future<?> monitor_future

bcast_task

protected final FD.Broadcaster bcast_task

Transmits SUSPECT message until view change or UNSUSPECT is received

Constructor Detail

FD

public FD()

Method Detail

getLocalAddress

public java.lang.String getLocalAddress()

getMembers

public java.lang.String getMembers()

getPingableMembers

public java.lang.String getPingableMembers()

getPingDest

public java.lang.String getPingDest()

getNumberOfHeartbeatsSent

public int getNumberOfHeartbeatsSent()

getNumSuspectEventsGenerated

public int getNumSuspectEventsGenerated()

getTimeout

public long getTimeout()

setTimeout

public void setTimeout(long timeout)

getMaxTries

public int getMaxTries()

setMaxTries

public void setMaxTries(int max_tries)

getCurrentNumTries

public int getCurrentNumTries()

printSuspectHistory

public java.lang.String printSuspectHistory()

resetStats

public void resetStats()

Overrides:

resetStats in class Protocol

init

public void init()
          throws java.lang.Exception

Description copied from class: Protocol

Called after instance has been created (null constructor) and before protocol is started. Properties are already set. Other protocols are not yet connected and events cannot yet be sent.

Overrides:

init in class Protocol

Throws:

java.lang.Exception - Thrown if protocol cannot be initialized successfully. This will cause the ProtocolStack to fail, so the channel constructor will throw an exception

stop

public void stop()

Description copied from class: Protocol

This method is called on a Channel.disconnect(). Stops work (e.g. by closing multicast socket). Will be called from top to bottom. This means that at the time of the method invocation the neighbor protocol below is still working. This method will replace the STOP, STOP_OK, CLEANUP and CLEANUP_OK events. The ProtocolStack guarantees that when this method is called all messages in the down queue will have been flushed

Overrides:

stop in class Protocol

getPingDest

protected Address getPingDest(java.util.List<Address> mbrs)

stopFailureDetection

public void stopFailureDetection()

startFailureDetection

public void startFailureDetection()

startMonitor

protected void startMonitor()

Requires lock to held by caller

stopMonitor

protected void stopMonitor()

Requires lock to be held by caller

isMonitorRunning

public boolean isMonitorRunning()

up

public java.lang.Object up(Event evt)

Description copied from class: Protocol

An event was received from the layer below. Usually the current layer will want to examine the event type and - depending on its type - perform some computation (e.g. removing headers from a MSG event type, or updating the internal membership list when receiving a VIEW_CHANGE event). Finally the event is either a) discarded, or b) an event is sent down the stack using down_prot.down() or c) the event (or another event) is sent up the stack using up_prot.up().

Overrides:

up in class Protocol

down

public java.lang.Object down(Event evt)

Description copied from class: Protocol

An event is to be sent down the stack. The layer may want to examine its type and perform some action on it, depending on the event's type. If the event is a message MSG, then the layer may need to add a header to it (or do nothing at all) before sending it down the stack using down_prot.down(). In case of a GET_ADDRESS event (which tries to retrieve the stack's address from one of the bottom layers), the layer may need to send a new response event back up the stack using up_prot.up().

Overrides:

down in class Protocol

sendHeartbeatResponse

protected void sendHeartbeatResponse(Address dest)

unsuspect

protected void unsuspect(Address mbr)

updateTimestamp

protected void updateTimestamp(Address sender)

computePingDest

protected void computePingDest(Address remove)

Computes pingable_mbrs (based on the current membership and the suspected members) and ping_dest

Parameters:

remove - The member to be removed from pingable_mbrs