Interface IDataStreamer<TK, TV>
Data streamer is responsible for loading external data into cache. It achieves it by properly buffering updates and properly mapping keys to nodes responsible for the data to make sure that there is the least amount of data movement possible and optimal network and memory utilization.
Note that streamer will load data concurrently by multiple internal threads, so the data may get to remote nodes in different order from which it was added to the streamer.
Also note that IDataStreamer
is not the only way to load data into cache.
Alternatively you can use
LoadCache(Action<TK, TV>, Object[])
method to load data from underlying data store. You can also use standard cache
put
and putAll
operations as well, but they most likely will not perform
as well as this class for loading data. And finally, data can be loaded from underlying
data store on demand, whenever it is accessed - for this no explicit data loading step
is needed.
IDataStreamer
supports the following configuration properties:
- PerNodeBufferSizeWhen entries are added to data streamer they are not sent to Ignite right away and are buffered internally for better performance and network utilization. This setting controls the size of internal per-node buffer before buffered data is sent to remote node. Default value is 1024.
- PerNodeParallelOperationsSometimes data may be added to the data streamer faster than it can be put in cache. In this case, new buffered load messages are sent to remote nodes before responses from previous ones are received. This could cause unlimited heap memory utilization growth on local and remote nodes. To control memory utilization, this setting limits maximum allowed number of parallel buffered load messages that are being processed on remote nodes. If this number is exceeded, then data streamer add/remove methods will block to control memory utilization. Default value is 16.
- AutoFlushFrequencyAutomatic flush frequency in milliseconds. Essentially, this is the time
after which the streamer will make an attempt to submit all data added so far to remote
nodes. Note that there is no guarantee that data will be delivered after this concrete
attempt (e.g., it can fail when topology is changing), but it won't be lost anyway.
Disabled by default (default value is
0
). - IsolatedDefines if data streamer will assume that there are no other concurrent
updates and allow data streamer choose most optimal concurrent implementation. Default value
is
false
.
All members are thread-safe and may be used concurrently from multiple threads.
Inherited Members
Namespace: Apache.Ignite.Core.Datastream
Assembly: Apache.Ignite.Core.dll
Syntax
public interface IDataStreamer<TK, TV> : IDisposable
Type Parameters
Name | Description |
---|---|
TK | |
TV |
Properties
AllowOverwrite
Gets or sets a value indicating whether existing values can be overwritten by the data streamer. Performance is better when this flag is false.
NOTE: When false, cache updates won't be propagated to cache store (even if SkipStore is false).
Default is false
.
Declaration
bool AllowOverwrite { get; set; }
Property Value
Type | Description |
---|---|
System.Boolean |
AutoFlushFrequency
Automatic flush frequency in milliseconds. Essentially, this is the time after which the streamer will make an attempt to submit all data added so far to remote nodes. Note that there is no guarantee that data will be delivered after this concrete attempt (e.g., it can fail when topology is changing), but it won't be lost anyway.
If set to 0
, automatic flush is disabled.
Default is 0
(disabled).
Declaration
long AutoFlushFrequency { get; set; }
Property Value
Type | Description |
---|---|
System.Int64 |
CacheName
Name of the cache to load data to.
Declaration
string CacheName { get; }
Property Value
Type | Description |
---|---|
System.String |
PerNodeBufferSize
Size of per node key-value pairs buffer.
Setter must be called before any add/remove operation.
Default is DefaultPerNodeBufferSize.
Declaration
int PerNodeBufferSize { get; set; }
Property Value
Type | Description |
---|---|
System.Int32 |
PerNodeParallelOperations
Maximum number of parallel load operations for a single node.
Setter must be called before any add/remove operation.
Default is 0, which means Ignite calculates this automatically as DataStreamerThreadPoolSize * DefaultParallelOperationsMultiplier.
Declaration
int PerNodeParallelOperations { get; set; }
Property Value
Type | Description |
---|---|
System.Int32 |
Receiver
Gets or sets custom stream receiver.
Declaration
IStreamReceiver<TK, TV> Receiver { get; set; }
Property Value
Type | Description |
---|---|
IStreamReceiver<TK, TV> |
SkipStore
Flag indicating that write-through behavior should be disabled for data loading.
AllowOverwrite must be true for write-through to work.
Default is false
.
Declaration
bool SkipStore { get; set; }
Property Value
Type | Description |
---|---|
System.Boolean |
Task
Gets the task for this loading process. This task completes whenever method Close(Boolean) completes.
Declaration
Task Task { get; }
Property Value
Type | Description |
---|---|
System.Threading.Tasks.Task |
Timeout
Gets or sets the timeout. Negative values mean no timeout. Default is DefaultTimeout.
Timeout is used in the following cases:
Declaration
TimeSpan Timeout { get; set; }
Property Value
Type | Description |
---|---|
System.TimeSpan |
Methods
AddData(TK, TV)
Adds single key-value pair for loading. Passing null
as value will be
interpreted as removal.
Declaration
Task AddData(TK key, TV val)
Parameters
Type | Name | Description |
---|---|---|
TK | key | Key. |
TV | val | Value. |
Returns
Type | Description |
---|---|
System.Threading.Tasks.Task | Task for this operation. |
AddData(ICollection<KeyValuePair<TK, TV>>)
Adds collection of key-value pairs for loading.
Declaration
Task AddData(ICollection<KeyValuePair<TK, TV>> entries)
Parameters
Type | Name | Description |
---|---|---|
System.Collections.Generic.ICollection<System.Collections.Generic.KeyValuePair<TK, TV>> | entries | Entries. |
Returns
Type | Description |
---|---|
System.Threading.Tasks.Task | Task for this operation. |
AddData(KeyValuePair<TK, TV>)
Adds single key-value pair for loading. Passing null
as pair's value will
be interpreted as removal.
Declaration
Task AddData(KeyValuePair<TK, TV> pair)
Parameters
Type | Name | Description |
---|---|---|
System.Collections.Generic.KeyValuePair<TK, TV> | pair | Key-value pair. |
Returns
Type | Description |
---|---|
System.Threading.Tasks.Task | Task for this operation. |
Close(Boolean)
Closes this streamer optionally loading any remaining data.
Declaration
void Close(bool cancel)
Parameters
Type | Name | Description |
---|---|---|
System.Boolean | cancel | Whether to cancel ongoing loading operations. When set to |
Flush()
Loads any remaining data, but doesn't close the streamer. Data can be still added after flush is finished. This method blocks and doesn't allow to add any data until all data is loaded.
Declaration
void Flush()
RemoveData(TK)
Adds key for removal.
Declaration
Task RemoveData(TK key)
Parameters
Type | Name | Description |
---|---|---|
TK | key | Key. |
Returns
Type | Description |
---|---|
System.Threading.Tasks.Task | Task for this operation. |
TryFlush()
Makes an attempt to load remaining data. This method is mostly similar to Flush() with the difference that it won't wait and will exit immediately.
Declaration
void TryFlush()
WithKeepBinary<TK1, TV1>()
Gets streamer instance with binary mode enabled, changing key and/or value types if necessary. In binary mode stream receiver gets data in binary format. You can only change key/value types when transitioning from non-binary to binary streamer; Changing type of binary streamer is not allowed and will throw an System.InvalidOperationException
Declaration
IDataStreamer<TK1, TV1> WithKeepBinary<TK1, TV1>()
Returns
Type | Description |
---|---|
IDataStreamer<TK1, TV1> | Streamer instance with binary mode enabled. |
Type Parameters
Name | Description |
---|---|
TK1 | Key type in binary mode. |
TV1 | Value type in binary mode. |