Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
56 changes: 53 additions & 3 deletions docs/streams/upgrade-guide.html
Original file line number Diff line number Diff line change
Expand Up @@ -71,14 +71,64 @@ <h1>Upgrade Guide and API Changes</h1>
</p>

<h3><a id="streams_api_changes_230" href="#streams_api_changes_230">Streams API changes in 2.3.0</a></h3>

<p>Version 2.3.0 adds the Suppress operator to the <code>kafka-streams-scala</code> Ktable API.</p>

<p>
As of 2.3.0 Streams now offers an in-memory version of the window (<a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-428%3A+Add+in-memory+window+store">KIP-428</a>)
and the session (<a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-445%3A+In-memory+Session+Store">KIP-445</a>) store, in addition to the persistent ones based on RocksDB.
The new public interfaces <code>inMemoryWindowStore()</code> and <code>inMemorySessionStore()</code> are added to <code>Stores</code> and provide the built-in in-memory window or session store.
</p>

<p>
In 2.3.0 we have added default implementation to <code>close()</code> and <code>configure()</code> for <code>Serializer</code>,
<code>Deserializer</code> and <code>Serde</code> so that they can be implemented by lambda expression.
For more details please read <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-331+Add+default+implementation+to+close%28%29+and+configure%28%29+for+Serializer%2C+Deserializer+and+Serde">KIP-331</a>.
</p>

<p>
To improve operator semantics, new store types are added that allow storing an additional timestamp per key-value pair or window.
Some DSL operators (for example KTables) are using those new stores.
Hence, you can now retrieve the last update timestamp via Interactive Queries if you specify
<code>TimestampedKeyValueStoreType</code> or <code>TimestampedWindowStoreType</code> as your <code>QueryableStoreType</code>.
While this change is mainly transparent, there are some corner cases that may require code changes:
<strong>Caution: If you receive an untyped store and use a cast, you might need to update your code to cast to the correct type.
Otherwise, you might get an exception similar to
<code>java.lang.ClassCastException: class org.apache.kafka.streams.state.ValueAndTimestamp cannot be cast to class YOUR-VALUE-TYPE</code>
upon getting a value from the store.</strong>
Additionally, <code>TopologyTestDriver#getStateStore()</code> only returns non-built-in stores and throws an exception if a built-in store is accessed.
For more details please read <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-258%3A+Allow+to+Store+Record+Timestamps+in+RocksDB">KIP-258</a>.
</p>

<p>
To improve type safety, a new operator <code>KStream#flatTransformValues</code> is added.
For more details please read <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-313%3A+Add+KStream.flatTransform+and+KStream.flatTransformValues">KIP-313</a>.
</p>

<p>
Kafka Streams used to set the configuration parameter <code>max.poll.interval.ms</code> to <code>Integer.MAX_VALUE</code>.
This default value is removed and Kafka Streams uses the consumer default value now.
For more details please read <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-442%3A+Return+to+default+max+poll+interval+in+Streams">KIP-442</a>.
</p>

<p>
Default configuration for repartition topic was changed:
The segment size for index files (<code>segment.index.bytes</code>) is no longer 50MB, but uses the cluster default.
Similarly, the configuration <code>segment.ms</code> in no longer 10 minutes, but uses the cluster default configuration.
Lastly, the retention period (<code>retention.ms</code>) is changed from <code>Long.MAX_VALUE</code> to <code>-1</code> (infinite).
For more details please read <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-443%3A+Return+to+default+segment.ms+and+segment.index.bytes+in+Streams+repartition+topics">KIP-443</a>.
</p>

<p>
As of 2.3.0 Streams now offers an in-memory version of the window and the session store, in addition to the persistent ones based on RocksDB. The new public interfaces <code>inMemoryWindowStore()</code> and <code>inMemorySessionStore()</code> are added to Stores and provide the built-in in-memory window or session store.
To avoid memory leaks, <code>RocksDBConfigSetter</code> has a new <code>close()</code> method that is called on shutdown.
Users should implement this method to release any memory used by RocksDB config objects, by closing those objects.
For more details please read <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-453%3A+Add+close%28%29+method+to+RocksDBConfigSetter">KIP-453</a>.
</p>

<p>
In 2.3.0 we have added default implementation to close() and configure() for <code>Serializer</code>, <code>Deserializer</code> and <code>Serde</code> so that they can be
implemented by lambda expression. For more details please read <a href="https://cwiki.apache.org/confluence/display/KAFKA/KIP-331+Add+default+implementation+to+close%28%29+and+configure%28%29+for+Serializer%2C+Deserializer+and+Serde">KIP-331</a>.
RocksDB dependency was updated to version <code>5.18.3</code>.
The new version allows to specify more RocksDB configurations, including <code>WriteBufferManager</code> that help to limit RocksDB off-head memory usage.
For more details please read <a href="https://issues.apache.org/jira/browse/KAFKA-8215">KAFKA-8215</a>.
</p>

<h3><a id="streams_api_changes_220" href="#streams_api_changes_220">Streams API changes in 2.2.0</a></h3>
Expand Down
10 changes: 10 additions & 0 deletions docs/upgrade.html
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,16 @@

<script id="upgrade-template" type="text/x-handlebars-template">

<h4><a id="upgrade_2_3_0" href="#upgrade_2_3_0">Upgrading from 0.8.x, 0.9.x, 0.10.0.x, 0.10.1.x, 0.10.2.x, 0.11.0.x, 1.0.x, 1.1.x, 2.0.x or 2.1.x or 2.2.x to 2.3.0</a></h4>

<!-- TODO core-team -->

<h5><a id="upgrade_230_notable" href="#upgrade_230_notable">Notable changes in 2.3.0</a></h5>
<ul>
<li> Kafka Streams DSL switches its used store types. While this change is mainly transparent to users, there are some corner cases that may require code changes.
See the <a href="/{{version}}/documentation/streams/upgrade-guide#streams_api_changes_230">Kafka Streams upgrade section</a> for more details.</li>
</ul>

<h4><a id="upgrade_2_2_0" href="#upgrade_2_2_0">Upgrading from 0.8.x, 0.9.x, 0.10.0.x, 0.10.1.x, 0.10.2.x, 0.11.0.x, 1.0.x, 1.1.x, 2.0.x or 2.1.x to 2.2.0</a></h4>

<p><b>If you are upgrading from a version prior to 2.1.x, please see the note below about the change to the schema used to store consumer offsets.
Expand Down