[GitHub] [flink] alpinegizmo commented on a change in pull request #10937: [FLINK-15743][docs] Add release notes for Flink 1.10

[GitBox <gi...@apache.org>]

alpinegizmo commented on a change in pull request #10937: [FLINK-15743][docs] Add release notes for Flink 1.10
URL: https://github.com/apache/flink/pull/10937#discussion_r370736220
 
 

 ##########
 File path: docs/release-notes/flink-1.10.md
 ##########
 @@ -0,0 +1,377 @@
+---
+title: "Release Notes - Flink 1.10"
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements.  See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership.  The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License.  You may obtain a copy of the License at
+
+  http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied.  See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+
+These release notes discuss important aspects, such as configuration, behavior,
+or dependencies, that changed between Flink 1.9 and Flink 1.10. Please read
+these notes carefully if you are planning to upgrade your Flink version to 1.10.
+
+* This will be replaced by the TOC
+{:toc}
+
+
+### Clusters & Deployment
+#### FileSystems should be loaded via Plugin Architecture ([FLINK-11956](https://issues.apache.org/jira/browse/FLINK-11956))
+In the s3-hadoop and s3-presto filesystems, classes from external
+dependencies, such as the AWS SDK, are no longer relocated. In the past, class
+relocation turned out to be problematic in combination with custom
+implementations of the `AWSCredentialsProvider` interface. As a consequence of
+removing class relocation, s3-hadoop and s3-presto filesystems can only be
+used as [plugins]({{ site.baseurl }}/ops/filesystems/#pluggable-file-systems).
+Other filesystems are strongly recommended to be only used as plugins.
+
+#### Flink Client respects Classloading Policy ([FLINK-13749](https://issues.apache.org/jira/browse/FLINK-13749))
+The Flink client now also respects the configured classloading policy, i.e.,
+`parent-first` or `child-first` classloading. Previously, only cluster
+components such as the job manager or task manager supported this setting.
+This does mean that users might get different behaviour in their programs, in
+which case they should configure the classloading policy explicitly to use
+`parent-first` classloading, which was the previous (hard-coded) behaviour.
+
+#### Enable spreading out Tasks evenly across all TaskManagers ([FLINK-12122](https://issues.apache.org/jira/browse/FLINK-12122))
+When [FLIP-6](https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=65147077)
+was rolled out with Flink 1.5.0, we changed how slots are allocated
+from TaskManagers (TMs). Instead of evenly allocating the slots from all
+registered TMs, we had the tendency to exhaust a TM before using another one.
+To use a scheduling strategy that is more similar to the pre-FLIP-6
+behaviour, where Flink tries to spread out the workload across all available
+TMs, one can set `cluster.evenly-spread-out-slots: true` in the
+`flink-conf.yaml`.
+
+#### Directory Structure Change for highly available Artifacts ([FLINK-13633](https://issues.apache.org/jira/browse/FLINK-13633))
+All highly available artifacts stored by Flink will now be stored under
+`HA_STORAGE_DIR/HA_CLUSTER_ID` with `HA_STORAGE_DIR` configured by
+`high-availability.storageDir` and `HA_CLUSTER_ID` configured by
+`high-availability.cluster-id`.
+
+#### Resources and JARs shipped via --yarnship will be ordered in the Classpath ([FLINK-13127](https://issues.apache.org/jira/browse/FLINK-13127))
+When using the `--yarnship` command line option, resource directories and jar
+files will be added to the classpath in lexicographical order with resources
+directories appearing first.
+
+#### Removal of --yn/--yarncontainer Command Line Options ([FLINK-12362](https://issues.apache.org/jira/browse/FLINK-12362))
+The Flink CLI no longer supports the deprecated command line options
+`-yn/--yarncontainer`, which were used to specify the number of containers to
+start on YARN. This option has been deprecated since the introduction of
+[FLIP-6](https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=65147077).
+All Flink users are advised to remove this command line option.
+
+#### Mesos Integration will reject expired Offers faster ([FLINK-14029](https://issues.apache.org/jira/browse/FLINK-14029))
+Flink's Mesos integration now rejects all expired offers instead of only 4.
+This improves the situation where Fenzo holds on to a lot of expired offers
+without giving them back to the Mesos resource manager.
+
+#### Scheduler Rearchitecture ([FLINK-14651](https://issues.apache.org/jira/browse/FLINK-14651))
+Flink's scheduler was refactored with the goal of making scheduling strategies
+customizable in the future. Users that experience issues related to scheduling
+can fallback to the legacy scheduler by setting `jobmanager.scheduler: legacy`
+in their `flink-conf.yaml`.
+
+
+### Memory Management
+#### New Task Executor Memory Model ([FLINK-13980](https://issues.apache.org/jira/browse/FLINK-13980))
+With
+[FLIP-49](https://cwiki.apache.org/confluence/display/FLINK/FLIP-49%3A+Unified+Memory+Configuration+for+TaskExecutors),
+a new memory model has been introduced for the task executor. New configuration
+options have been introduced to control the memory consumption of the task
+executor process. This affects all types of deployments: standalone, YARN,
+Mesos, and the new active Kubernetes integration. The memory model of the job
+manager process has not been changed yet but it is planned to be updated as
+well.
+
+If you try to reuse your previous Flink configuration without any adjustments,
+the new memory model can result in differently computed memory parameters for
+the JVM and, thus, performance changes.
+
+Please check the user documentation <!-- TODO: insert link --> for more details.
+
+##### Deprecation and breaking changes
+The following options have been removed and have no effect anymore:
+
+<table class="table">
+  <thead>
+    <tr>
+      <th class="text-left" style="width: 30%">Deprecated/removed config option</th>
+      <th class="text-left">Note</th>
+    </tr>
+  </thead>
+  <tbody>
+    <tr>
+      <td>taskmanager.memory.fraction</td>
+      <td>
+        Check also the description of the new option
+        <code class="highlighter-rouge">taskmanager.memory.managed.fraction</code>
+        but it has different semantics and the value of the deprecated option
+        usually has to be adjusted
+      </td>
+    </tr>
+    <tr>
+      <td>taskmanager.memory.off-heap</td>
+      <td>On-heap managed memory is no longer supported</td>
 
 Review comment:
   Which is it that's no longer supported: off-heap or on-heap?

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services