com.aquenos.cassandra.pvarchiver.server.archiving

Class ArchivingService

java.lang.Object

com.aquenos.cassandra.pvarchiver.server.archiving.ArchivingService

All Implemented Interfaces:

DisposableBean, InitializingBean, SmartInitializingSingleton, SelfNaming

@ManagedResource(description="handles archiving operations")
public class ArchivingService
extends Object
implements DisposableBean, InitializingBean, SelfNaming, SmartInitializingSingleton

Service performing the actual archiving. This service gets the list of channels registered with this server when the server goes online and starts archiving for these channels. It stops archiving when the server goes offline. The configuration of individual channels can be reloaded using the refreshChannel(String) method.

Author:

Sebastian Marsching

Constructor Summary

Constructors

Constructor and Description
ArchivingService()

Method Summary

Modifier and Type	Method and Description
void	afterPropertiesSet()
void	afterSingletonsInstantiated()
void	destroy()
ChannelStatus	getChannelStatus(String channelName) Returns the status of an archived channel.
List<ChannelStatus>	getChannelStatusForAllChannels() Returns a list of the status for all channels known by this archiving service.
long	getNumberOfSamplesDropped() Returns the total number of samples that have been dropped.
long	getNumberOfSamplesWritten() Returns the total number of samples that have been written.
ObjectName	getObjectName()
int	getSamplesDecimationCurrentRunningFetchOperations() Returns the number of sample fetch operations that are running currently.
int	getSamplesDecimationCurrentSamplesInMemory() Returns the number of samples that have been fetched (but not processed yet).
int	getSamplesDecimationMaxRunningFetchOperations() Returns the max. number of sample fetch operations that may run concurrently.
int	getSamplesDecimationMaxSamplesInMemory() Returns the max. number of samples that may concurrently be kept in memory.
void	onServerOnlineStatusEvent(ServerOnlineStatusEvent event) Notifies the archiving service that the server's online status changed.
ListenableFuture<Void>	refreshChannel(String channelName) Reloads the configuration of the specified channel.
void	setArchiveAccessService(ArchiveAccessService archiveAccessService) Sets the archive access service that is used to read samples from the database.
void	setCassandraProvider(CassandraProvider cassandraProvider) Sets the Cassandra provider that provides access to the Apache Cassandra database.
void	setChannelMetaDataDAO(ChannelMetaDataDAO channelMetaDataDAO) Sets the DAO for reading and modifying meta-data related to channels.
void	setClusterManagementService(ClusterManagementService clusterManagementService) Sets the cluster management service.
void	setControlSystemSupportRegistry(ControlSystemSupportRegistry controlSystemSupportRegistry) Sets the control-system support registry.
void	setServerProperties(ServerProperties serverProperties) Sets the server-specific configuration properties.
void	setThrottlingProperties(ThrottlingProperties throttlingProperties) Sets the configuration properties controlling throttling.

Methods inherited from class java.lang.Object

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

Constructor Detail

ArchivingService

public ArchivingService()

Method Detail

afterPropertiesSet

public void afterPropertiesSet()
                        throws Exception

Specified by:

afterPropertiesSet in interface InitializingBean

Throws:

Exception

afterSingletonsInstantiated

public void afterSingletonsInstantiated()

Specified by:

afterSingletonsInstantiated in interface SmartInitializingSingleton

destroy

public void destroy()
             throws Exception

Specified by:

destroy in interface DisposableBean

Throws:

Exception

getObjectName

public ObjectName getObjectName()
                         throws MalformedObjectNameException

Specified by:

getObjectName in interface SelfNaming

Throws:

MalformedObjectNameException

getChannelStatus

public ChannelStatus getChannelStatus(String channelName)

Returns the status of an archived channel. If the channel is not known by this archiving service, null is returned. Throws an IllegalStateException if the archiving service has not finished initialization of its channel list yet.

Parameters:

channelName - name of the channel to be queried.

Returns:

status of the specified channel or null if the specified channel is not known by this archiving service.

Throws:

IllegalStateException - if the archiving service has not initialized its list of channels yet.

getChannelStatusForAllChannels

public List<ChannelStatus> getChannelStatusForAllChannels()

Returns a list of the status for all channels known by this archiving service. Throws an IllegalStateException if the archiving service has not finished initialization of its channel list yet. The channels are returned in the natural order of their names. The returned list is never null, but it might be empty if no channels belong this server.

Returns:

status of all channels known by this archiving service.

Throws:

IllegalStateException - if the archiving service has not initialized its list f channels yet.

getNumberOfSamplesDropped

@ManagedAttribute(description="number of samples that been dropped since the archiving service was started")
public long getNumberOfSamplesDropped()

Returns the total number of samples that have been dropped. A sample is dropped when new samples for a channel arrive more quickly than they can be written and consequently samples remain in the queue until a timeout is reached. The returned number represents the total number of samples that have been dropped for any channel since this archiving service was started. When this number grows constantly, this can be an indication that the server (or the Cassandra cluster) is overloaded and the workload should be reduced (possibly by adding more servers).

Returns:

number of samples that have been dropped since this archiving service was started.

getNumberOfSamplesWritten

@ManagedAttribute(description="number of samples that been writen since the archiving service was started")
public long getNumberOfSamplesWritten()

Returns the total number of samples that have been written. A sample is only counted as written when the databases acknowledges that the write operation has been successful. Samples for all decimation levels and not just raw samples contribute to this number. The returned number represents the total number of samples that have been written since this archiving service was started.

Returns:

number of samples that have been written since this archiving service was started.

getSamplesDecimationCurrentRunningFetchOperations

@ManagedAttribute(description="number of sample fetch operations (to generate decimated samples) that are currently running")
public int getSamplesDecimationCurrentRunningFetchOperations()

Returns the number of sample fetch operations that are running currently.

Usually, decimated samples are generated as new (raw) samples are written. However, when adding a new decimation level or when restarting the server, it might be necessary to read samples from the database in order to generate decimated samples. In this case, the number of fetch operations that may run concurrently is limited (to the number returned by getSamplesDecimationMaxRunningFetchOperations()).

Returns:

number of sample fetch operations (to generate decimated samples) that are currently running.

See Also:

getSamplesDecimationCurrentSamplesInMemory(), getSamplesDecimationMaxRunningFetchOperations()

getSamplesDecimationCurrentSamplesInMemory

@ManagedAttribute(description="current number of samples that have been fetched (to generate decimated samples), but not processed yet")
public int getSamplesDecimationCurrentSamplesInMemory()

Returns the number of samples that have been fetched (but not processed yet). This is the number of samples that is currently kept in memory.

Usually, decimated samples are generated as new (raw) samples are written. However, when adding a new decimation level or when restarting the server, it might be necessary to read samples from the database in order to generate decimated samples. For performance reasons, samples are fetched in batches. These samples, that have been fetched, but not processed yet, consume memory, so that no new fetch operations may be started if a large number of samples has already been fetched, but not processed yet. The actual limit is returned by getSamplesDecimationMaxSamplesInMemory().

Returns:

number of samples that have been fetched (to generate decimated samples), but not processed yet.

See Also:

getSamplesDecimationCurrentRunningFetchOperations(), getSamplesDecimationMaxSamplesInMemory()

getSamplesDecimationMaxRunningFetchOperations

@ManagedAttribute(description="max. number of sample fetch operations (to generate decimated samples) that may run concurrently")
public int getSamplesDecimationMaxRunningFetchOperations()

Returns the max. number of sample fetch operations that may run concurrently.

Usually, decimated samples are generated as new (raw) samples are written. However, when adding a new decimation level or when restarting the server, it might be necessary to read samples from the database in order to generate decimated samples. In this case, the number of fetch operations that may run concurrently is limited to the number returned by this method.

Returns:

max. number of sample fetch operations (to generate decimated samples) that may run concurrently.

See Also:

getSamplesDecimationCurrentRunningFetchOperations(), getSamplesDecimationMaxSamplesInMemory()

getSamplesDecimationMaxSamplesInMemory

@ManagedAttribute(description="max. number of samples that may be fetched into memory")
public int getSamplesDecimationMaxSamplesInMemory()

Returns the max. number of samples that may concurrently be kept in memory.

Usually, decimated samples are generated as new (raw) samples are written. However, when adding a new decimation level or when restarting the server, it might be necessary to read samples from the database in order to generate decimated samples. For performance reasons, samples are fetched in batches. These samples, that have been fetched, but not processed yet, consume memory, so that no new fetch operations may be started if a large number of samples has already been fetched, but not processed yet. This method returns the number of samples at which this limit is enforced and no more samples are fetched.

Returns:

max. number of samples that may be fetched (to generate decimated samples) into memory concurrently.

See Also:

getSamplesDecimationCurrentSamplesInMemory(), getSamplesDecimationMaxRunningFetchOperations()

onServerOnlineStatusEvent

@EventListener
public void onServerOnlineStatusEvent(ServerOnlineStatusEvent event)

Notifies the archiving service that the server's online status changed. Typically, this method is automatically called by the Spring container when the ClusterManagementService publishes ServerOnlineStatusEvent.

This method takes care of starting archiving when the server goes online and stopping it when the server goes offline. These operations are performed asynchronously so that the thread sending the notification does not block.

Parameters:

event - event specifying the server's new online status.

refreshChannel

public ListenableFuture<Void> refreshChannel(String channelName)

Reloads the configuration of the specified channel. This method should be called whenever the configuration of a channel belonging to this server has been changed. It also needs to be called after registering a pending operation for a channel, so that archiving for the channel can be disabled.

If the server is currently offline, calling this method will typically have no effect because there is no need to reload the channel configuration. In this case, the future returned by this method completed immediately.

The future returned by this method completes when the refresh operation has finished. It should never throw an exception.

Parameters:

channelName - name of the channel to be refreshed.

Returns:

future that completes when the refresh operation has finished.

Throws:

NullPointerException - if channelName is null.

setArchiveAccessService

@Autowired
public void setArchiveAccessService(ArchiveAccessService archiveAccessService)

Sets the archive access service that is used to read samples from the database. When generating decimated samples, the archiving service might have to read samples that have been written earlier. It uses the archive access service for this purpose. Typically, this method is called automatically by the Spring container.

Parameters:

archiveAccessService - archive access service used for reading samples when generating decimated samples.

setCassandraProvider

@Autowired
public void setCassandraProvider(CassandraProvider cassandraProvider)

Sets the Cassandra provider that provides access to the Apache Cassandra database. Typically, this method is called automatically by the Spring container.

Parameters:

cassandraProvider - provider that provides a connection to the Apache Cassandra database.

setChannelMetaDataDAO

@Autowired
public void setChannelMetaDataDAO(ChannelMetaDataDAO channelMetaDataDAO)

Sets the DAO for reading and modifying meta-data related to channels. The archiving services uses this DAO to read the configuration for archived channels and create the meta-data for new sample buckets when needed. Typically, this method is called automatically by the Spring container.

Parameters:

channelMetaDataDAO - channel meta-data DAO to be used by this object.

setClusterManagementService

@Autowired
public void setClusterManagementService(ClusterManagementService clusterManagementService)

Sets the cluster management service. The cluster management service is used to get information about the cluster status. In particular, it is needed to determine the online state of this server in order to now when archiving should be started and stopped. Typically, this method is called automatically by the Spring container.

Parameters:

clusterManagementService - cluster manager used by the archiving server.

setControlSystemSupportRegistry

@Autowired
public void setControlSystemSupportRegistry(ControlSystemSupportRegistry controlSystemSupportRegistry)

Sets the control-system support registry. The registry is used to gain access to the control-system supports available on this archiving server. Typically, this method is called automatically by the Spring container.

Parameters:

controlSystemSupportRegistry - control-system support registry for this archiving server.

setThrottlingProperties

@Autowired
public void setThrottlingProperties(ThrottlingProperties throttlingProperties)

Sets the configuration properties controlling throttling. These configuration properties are used to limit resource usage by the sample decimation process..

Parameters:

throttlingProperties - configuration properties controlling throttling.

setServerProperties

@Autowired
public void setServerProperties(ServerProperties serverProperties)

Sets the server-specific configuration properties. These server properties are used to get the ID of this archiving server.

Parameters:

serverProperties - configuration properties used by this server.