HyukjinKwon commented on a change in pull request #31598:
URL: https://github.com/apache/spark/pull/31598#discussion_r584449917
##########
File path: docs/configuration.md
##########
@@ -114,12 +114,67 @@ in the `spark-defaults.conf` file. A few configuration
keys have been renamed si
versions of Spark; in such cases, the older key names are still accepted, but
take lower
precedence than any instance of the newer key.
-Spark properties mainly can be divided into two kinds: one is related to
deploy, like
-"spark.driver.memory", "spark.executor.instances", this kind of properties may
not be affected when
-setting programmatically through `SparkConf` in runtime, or the behavior is
depending on which
-cluster manager and deploy mode you choose, so it would be suggested to set
through configuration
-file or `spark-submit` command line options; another is mainly related to
Spark runtime control,
-like "spark.task.maxFailures", this kind of properties can be set in either
way.
+Note that Spark properties have different effective timing and they can be
divided into three kinds:
+<table class="table">
+<tr><th>Configuration Type</th><th>Meaning</th><th>Examples</th></tr>
+<tr>
+ <td><code>Configurations needed at driver launch</code></td>
+ <td>
+ Configuration used to submit an application, such as
<code>spark.driver.memory</code>, <code>spark.driver.extraClassPath</code>,
these kind of properties only effect before driver's JVM is started, so it
would be suggested to set through configuration file or
<code>spark-submit</code> command line options.
+ </td>
+ <td>
+ The following is a list of such configurations:
+ <ul>
+ <li><code>spark.driver.memory</code></li>
+ <li><code>spark.driver.memoryOverhead</code></li>
+ <li><code>spark.driver.cores</code></li>
+ <li><code>spark.driver.userClassPathFirst</code></li>
+ <li><code>spark.driver.extraClassPath</code></li>
+ <li><code>spark.driver.defaultJavaOptions</code></li>
+ <li><code>spark.driver.extraJavaOptions</code></li>
+ <li><code>spark.driver.extraLibraryPath</code></li>
+ <li><code>spark.driver.resource.*</code></li>
+ <li><code>spark.pyspark.driver.python</code></li>
+ <li><code>spark.pyspark.python</code></li>
+ <li><code>spark.r.shell.command</code></li>
+ <li><code>spark.launcher.childProcLoggerName</code></li>
+ <li><code>spark.launcher.childConnectionTimeout</code></li>
+ <li><code>spark.yarn.driver.*</code></li>
+ </ul>
+ </td>
+</tr>
+<tr>
+ <td><code>Application Deploy Related Configuration</code></td>
+ <td>
+ Like <code>spark.master</code>, <code>spark.executor.instances</code>,
this kind of properties may not be affected when setting programmatically
through <code>SparkConf</code> in runtime after SparkContext has been started,
or the behavior is depending on which cluster manager and deploy mode you
choose, so it would be suggested to set through configuration file,
<code>spark-submit</code> command line options, or setting programmatically
through <code>SparkConf</code> in runtime before start SparkContext.
+ </td>
+ <td>
+ The following is a example such configurations:
+ <ul>
+ <li><code>spark.master</code></li>
+ <li><code>spark.app.name</code></li>
+ <li><code>spark.executor.memory</code></li>
+ <li><code>spark.submit.deployMode</code></li>
+ <li><code>spark.eventLog.enabled</code></li>
+ <li><code>etc...</code></li>
+ </ul>
+ </td>
+</tr>
+<tr>
+ <td><code>Spark Runtime Control Related Configuration</code></td>
Review comment:
Or even we can just remove this section.
----------------------------------------------------------------
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:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
