This is an automated email from the ASF dual-hosted git repository.
dongjoon pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/spark.git
The following commit(s) were added to refs/heads/master by this push:
new 2f3cb36cb99 [SPARK-45261][CORE] Fix `EventLogFileWriters` to handle
`none` as a codec
2f3cb36cb99 is described below
commit 2f3cb36cb99a4b8d1ddec74696c4ed036c5df5b2
Author: Dongjoon Hyun <[email protected]>
AuthorDate: Thu Sep 21 14:44:11 2023 -0700
[SPARK-45261][CORE] Fix `EventLogFileWriters` to handle `none` as a codec
### What changes were proposed in this pull request?
This PR aims to support `none` as a codec instead of throwing exception.
Currrently, our unit test is supposed to test it, but actually it's not
tested at all.
https://github.com/apache/spark/blob/892fdc532696e703b353c4758320d69162fffe8c/core/src/test/scala/org/apache/spark/deploy/history/EventLogFileReadersSuite.scala#L120-L124
```
$ build/sbt "core/testOnly *EventLogFileReaderSuite*"
...
[info] - get information, list event log files, zip log files - with codec
None (33 milliseconds)
[info] - get information, list event log files, zip log files - with codec
Some(lz4) (125 milliseconds)
...
```
### Why are the changes needed?
```
$ bin/spark-shell \
-c spark.eventLog.enabled=true \
-c spark.eventLog.compress=true \
-c spark.eventLog.compression.codec=none
...
23/09/21 13:26:45 ERROR SparkContext: Error initializing SparkContext.
org.apache.spark.SparkIllegalArgumentException:
[CODEC_SHORT_NAME_NOT_FOUND] Cannot find a short name for the codec none.
```
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Pass the CIs with the revised CIs.
### Was this patch authored or co-authored using generative AI tooling?
No.
Closes #43038 from dongjoon-hyun/SPARK-45261.
Authored-by: Dongjoon Hyun <[email protected]>
Signed-off-by: Dongjoon Hyun <[email protected]>
---
core/src/main/scala/org/apache/spark/SparkContext.scala | 3 ++-
.../org/apache/spark/deploy/history/EventLogFileWriters.scala | 3 ++-
core/src/main/scala/org/apache/spark/internal/config/package.scala | 1 +
.../scala/org/apache/spark/deploy/history/EventLogTestHelper.scala | 7 ++++---
4 files changed, 9 insertions(+), 5 deletions(-)
diff --git a/core/src/main/scala/org/apache/spark/SparkContext.scala
b/core/src/main/scala/org/apache/spark/SparkContext.scala
index b90601a5bbb..39c8e483ebd 100644
--- a/core/src/main/scala/org/apache/spark/SparkContext.scala
+++ b/core/src/main/scala/org/apache/spark/SparkContext.scala
@@ -464,7 +464,8 @@ class SparkContext(config: SparkConf) extends Logging {
}
_eventLogCodec = {
- val compress = _conf.get(EVENT_LOG_COMPRESS)
+ val compress = _conf.get(EVENT_LOG_COMPRESS) &&
+ !_conf.get(EVENT_LOG_COMPRESSION_CODEC).equals("none")
if (compress && isEventLogEnabled) {
Some(_conf.get(EVENT_LOG_COMPRESSION_CODEC)).map(CompressionCodec.getShortName)
} else {
diff --git
a/core/src/main/scala/org/apache/spark/deploy/history/EventLogFileWriters.scala
b/core/src/main/scala/org/apache/spark/deploy/history/EventLogFileWriters.scala
index 7d44cbd9f64..418d9171842 100644
---
a/core/src/main/scala/org/apache/spark/deploy/history/EventLogFileWriters.scala
+++
b/core/src/main/scala/org/apache/spark/deploy/history/EventLogFileWriters.scala
@@ -54,7 +54,8 @@ abstract class EventLogFileWriter(
sparkConf: SparkConf,
hadoopConf: Configuration) extends Logging {
- protected val shouldCompress = sparkConf.get(EVENT_LOG_COMPRESS)
+ protected val shouldCompress = sparkConf.get(EVENT_LOG_COMPRESS) &&
+ !sparkConf.get(EVENT_LOG_COMPRESSION_CODEC).equals("none")
protected val shouldOverwrite = sparkConf.get(EVENT_LOG_OVERWRITE)
protected val outputBufferSize =
sparkConf.get(EVENT_LOG_OUTPUT_BUFFER_SIZE).toInt
protected val fileSystem = Utils.getHadoopFileSystem(logBaseDir, hadoopConf)
diff --git a/core/src/main/scala/org/apache/spark/internal/config/package.scala
b/core/src/main/scala/org/apache/spark/internal/config/package.scala
index 05b2624b403..3da61f6c81d 100644
--- a/core/src/main/scala/org/apache/spark/internal/config/package.scala
+++ b/core/src/main/scala/org/apache/spark/internal/config/package.scala
@@ -1907,6 +1907,7 @@ package object config {
"the codec.")
.version("3.0.0")
.stringConf
+ .transform(_.toLowerCase(Locale.ROOT))
.createWithDefault("zstd")
private[spark] val BUFFER_SIZE =
diff --git
a/core/src/test/scala/org/apache/spark/deploy/history/EventLogTestHelper.scala
b/core/src/test/scala/org/apache/spark/deploy/history/EventLogTestHelper.scala
index a68086256d1..ea8da010859 100644
---
a/core/src/test/scala/org/apache/spark/deploy/history/EventLogTestHelper.scala
+++
b/core/src/test/scala/org/apache/spark/deploy/history/EventLogTestHelper.scala
@@ -41,9 +41,10 @@ object EventLogTestHelper {
conf.set(EVENT_LOG_BLOCK_UPDATES, true)
conf.set(EVENT_LOG_TESTING, true)
conf.set(EVENT_LOG_DIR, logDir.toString)
- compressionCodec.foreach { codec =>
- conf.set(EVENT_LOG_COMPRESS, true)
- conf.set(EVENT_LOG_COMPRESSION_CODEC, codec)
+ conf.set(EVENT_LOG_COMPRESS, true)
+ compressionCodec match {
+ case Some(codec) => conf.set(EVENT_LOG_COMPRESSION_CODEC, codec)
+ case _ => conf.set(EVENT_LOG_COMPRESSION_CODEC, "None")
}
conf.set(EVENT_LOG_STAGE_EXECUTOR_METRICS, true)
conf
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
