[GitHub] [incubator-hudi] vinothchandar commented on a change in pull request #1418: [HUDI-678] Make config package spark free
vinothchandar commented on a change in pull request #1418: [HUDI-678] Make
config package spark free
URL: https://github.com/apache/incubator-hudi/pull/1418#discussion_r398317120
##
File path:
hudi-client/src/main/java/org/apache/hudi/config/HoodieMemoryConfig.java
##
@@ -113,52 +112,20 @@ public Builder withWriteStatusFailureFraction(double
failureFraction) {
return this;
}
-/**
- * Dynamic calculation of max memory to use for for spillable map.
user.available.memory = spark.executor.memory *
- * (1 - spark.memory.fraction) spillable.available.memory =
user.available.memory * hoodie.memory.fraction. Anytime
- * the spark.executor.memory or the spark.memory.fraction is changed, the
memory used for spillable map changes
- * accordingly
- */
-private long getMaxMemoryAllowedForMerge(String maxMemoryFraction) {
- final String SPARK_EXECUTOR_MEMORY_PROP = "spark.executor.memory";
- final String SPARK_EXECUTOR_MEMORY_FRACTION_PROP =
"spark.memory.fraction";
- // This is hard-coded in spark code {@link
- //
https://github.com/apache/spark/blob/576c43fb4226e4efa12189b41c3bc862019862c6/core/src/main/scala/org/apache/
- // spark/memory/UnifiedMemoryManager.scala#L231} so have to re-define
this here
- final String DEFAULT_SPARK_EXECUTOR_MEMORY_FRACTION = "0.6";
- // This is hard-coded in spark code {@link
- //
https://github.com/apache/spark/blob/576c43fb4226e4efa12189b41c3bc862019862c6/core/src/main/scala/org/apache/
- // spark/SparkContext.scala#L471} so have to re-define this here
- final String DEFAULT_SPARK_EXECUTOR_MEMORY_MB = "1024"; // in MB
-
- if (SparkEnv.get() != null) {
-// 1 GB is the default conf used by Spark, look at SparkContext.scala
-long executorMemoryInBytes = Utils.memoryStringToMb(
-SparkEnv.get().conf().get(SPARK_EXECUTOR_MEMORY_PROP,
DEFAULT_SPARK_EXECUTOR_MEMORY_MB)) * 1024 * 1024L;
-// 0.6 is the default value used by Spark,
-// look at {@link
-//
https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/SparkConf.scala#L507}
-double memoryFraction = Double.parseDouble(
-SparkEnv.get().conf().get(SPARK_EXECUTOR_MEMORY_FRACTION_PROP,
DEFAULT_SPARK_EXECUTOR_MEMORY_FRACTION));
-double maxMemoryFractionForMerge =
Double.parseDouble(maxMemoryFraction);
-double userAvailableMemory = executorMemoryInBytes * (1 -
memoryFraction);
-long maxMemoryForMerge = (long) Math.floor(userAvailableMemory *
maxMemoryFractionForMerge);
-return Math.max(DEFAULT_MIN_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES,
maxMemoryForMerge);
- } else {
-return DEFAULT_MAX_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES;
- }
-}
-
public HoodieMemoryConfig build() {
HoodieMemoryConfig config = new HoodieMemoryConfig(props);
setDefaultOnCondition(props,
!props.containsKey(MAX_MEMORY_FRACTION_FOR_COMPACTION_PROP),
MAX_MEMORY_FRACTION_FOR_COMPACTION_PROP,
DEFAULT_MAX_MEMORY_FRACTION_FOR_COMPACTION);
setDefaultOnCondition(props,
!props.containsKey(MAX_MEMORY_FRACTION_FOR_MERGE_PROP),
MAX_MEMORY_FRACTION_FOR_MERGE_PROP,
DEFAULT_MAX_MEMORY_FRACTION_FOR_MERGE);
+ long maxMemoryAllowedForMerge =
+
SparkConfigUtils.getMaxMemoryAllowedForMerge(props.getProperty(MAX_MEMORY_FRACTION_FOR_MERGE_PROP));
setDefaultOnCondition(props,
!props.containsKey(MAX_MEMORY_FOR_MERGE_PROP), MAX_MEMORY_FOR_MERGE_PROP,
-
String.valueOf(getMaxMemoryAllowedForMerge(props.getProperty(MAX_MEMORY_FRACTION_FOR_MERGE_PROP;
+ String.valueOf(maxMemoryAllowedForMerge));
+ long maxMemoryAllowedForCompaction =
+
SparkConfigUtils.getMaxMemoryAllowedForMerge(props.getProperty(MAX_MEMORY_FRACTION_FOR_COMPACTION_PROP));
Review comment:
if the classes in `config` call the SparkConfigUtils, then we cannot claim
its spark free right..
cc @yanghua as well
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:
us...@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-hudi] vinothchandar commented on a change in pull request #1418: [HUDI-678] Make config package spark free
vinothchandar commented on a change in pull request #1418: [HUDI-678] Make
config package spark free
URL: https://github.com/apache/incubator-hudi/pull/1418#discussion_r398316910
##
File path:
hudi-client/src/main/java/org/apache/hudi/config/HoodieMemoryConfig.java
##
@@ -113,40 +112,8 @@ public Builder withWriteStatusFailureFraction(double
failureFraction) {
return this;
}
-/**
- * Dynamic calculation of max memory to use for for spillable map.
user.available.memory = spark.executor.memory *
- * (1 - spark.memory.fraction) spillable.available.memory =
user.available.memory * hoodie.memory.fraction. Anytime
- * the spark.executor.memory or the spark.memory.fraction is changed, the
memory used for spillable map changes
- * accordingly
- */
private long getMaxMemoryAllowedForMerge(String maxMemoryFraction) {
- final String SPARK_EXECUTOR_MEMORY_PROP = "spark.executor.memory";
- final String SPARK_EXECUTOR_MEMORY_FRACTION_PROP =
"spark.memory.fraction";
- // This is hard-coded in spark code {@link
- //
https://github.com/apache/spark/blob/576c43fb4226e4efa12189b41c3bc862019862c6/core/src/main/scala/org/apache/
- // spark/memory/UnifiedMemoryManager.scala#L231} so have to re-define
this here
- final String DEFAULT_SPARK_EXECUTOR_MEMORY_FRACTION = "0.6";
- // This is hard-coded in spark code {@link
- //
https://github.com/apache/spark/blob/576c43fb4226e4efa12189b41c3bc862019862c6/core/src/main/scala/org/apache/
- // spark/SparkContext.scala#L471} so have to re-define this here
- final String DEFAULT_SPARK_EXECUTOR_MEMORY_MB = "1024"; // in MB
-
- if (SparkEnv.get() != null) {
-// 1 GB is the default conf used by Spark, look at SparkContext.scala
-long executorMemoryInBytes = Utils.memoryStringToMb(
-SparkEnv.get().conf().get(SPARK_EXECUTOR_MEMORY_PROP,
DEFAULT_SPARK_EXECUTOR_MEMORY_MB)) * 1024 * 1024L;
-// 0.6 is the default value used by Spark,
-// look at {@link
-//
https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/SparkConf.scala#L507}
-double memoryFraction = Double.parseDouble(
-SparkEnv.get().conf().get(SPARK_EXECUTOR_MEMORY_FRACTION_PROP,
DEFAULT_SPARK_EXECUTOR_MEMORY_FRACTION));
-double maxMemoryFractionForMerge =
Double.parseDouble(maxMemoryFraction);
-double userAvailableMemory = executorMemoryInBytes * (1 -
memoryFraction);
-long maxMemoryForMerge = (long) Math.floor(userAvailableMemory *
maxMemoryFractionForMerge);
-return Math.max(DEFAULT_MIN_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES,
maxMemoryForMerge);
- } else {
-return DEFAULT_MAX_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES;
- }
+ return ConfigUtils.getMaxMemoryAllowedForMerge(props, maxMemoryFraction);
Review comment:
@leesf I think this is still not addressed.. ?
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:
us...@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-hudi] vinothchandar commented on a change in pull request #1418: [HUDI-678] Make config package spark free
vinothchandar commented on a change in pull request #1418: [HUDI-678] Make
config package spark free
URL: https://github.com/apache/incubator-hudi/pull/1418#discussion_r398216210
##
File path:
hudi-client/src/main/java/org/apache/hudi/client/utils/ConfigUtils.java
##
@@ -0,0 +1,80 @@
+/*
+ * 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.
+ */
+
+package org.apache.hudi.client.utils;
+
+import org.apache.hudi.config.HoodieIndexConfig;
+
+import org.apache.spark.SparkEnv;
+import org.apache.spark.storage.StorageLevel;
+import org.apache.spark.util.Utils;
+
+import java.util.Properties;
+
+import static
org.apache.hudi.config.HoodieMemoryConfig.DEFAULT_MAX_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES;
+import static
org.apache.hudi.config.HoodieMemoryConfig.DEFAULT_MIN_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES;
+import static
org.apache.hudi.config.HoodieWriteConfig.WRITE_STATUS_STORAGE_LEVEL;
+
+/**
+ * Config utils.
+ */
+public class ConfigUtils {
Review comment:
rename to `SparkConfigUtils`? to make it explicit?
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:
us...@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-hudi] vinothchandar commented on a change in pull request #1418: [HUDI-678] Make config package spark free
vinothchandar commented on a change in pull request #1418: [HUDI-678] Make
config package spark free
URL: https://github.com/apache/incubator-hudi/pull/1418#discussion_r398217268
##
File path:
hudi-client/src/main/java/org/apache/hudi/config/HoodieMemoryConfig.java
##
@@ -113,40 +112,8 @@ public Builder withWriteStatusFailureFraction(double
failureFraction) {
return this;
}
-/**
- * Dynamic calculation of max memory to use for for spillable map.
user.available.memory = spark.executor.memory *
- * (1 - spark.memory.fraction) spillable.available.memory =
user.available.memory * hoodie.memory.fraction. Anytime
- * the spark.executor.memory or the spark.memory.fraction is changed, the
memory used for spillable map changes
- * accordingly
- */
private long getMaxMemoryAllowedForMerge(String maxMemoryFraction) {
- final String SPARK_EXECUTOR_MEMORY_PROP = "spark.executor.memory";
- final String SPARK_EXECUTOR_MEMORY_FRACTION_PROP =
"spark.memory.fraction";
- // This is hard-coded in spark code {@link
- //
https://github.com/apache/spark/blob/576c43fb4226e4efa12189b41c3bc862019862c6/core/src/main/scala/org/apache/
- // spark/memory/UnifiedMemoryManager.scala#L231} so have to re-define
this here
- final String DEFAULT_SPARK_EXECUTOR_MEMORY_FRACTION = "0.6";
- // This is hard-coded in spark code {@link
- //
https://github.com/apache/spark/blob/576c43fb4226e4efa12189b41c3bc862019862c6/core/src/main/scala/org/apache/
- // spark/SparkContext.scala#L471} so have to re-define this here
- final String DEFAULT_SPARK_EXECUTOR_MEMORY_MB = "1024"; // in MB
-
- if (SparkEnv.get() != null) {
-// 1 GB is the default conf used by Spark, look at SparkContext.scala
-long executorMemoryInBytes = Utils.memoryStringToMb(
-SparkEnv.get().conf().get(SPARK_EXECUTOR_MEMORY_PROP,
DEFAULT_SPARK_EXECUTOR_MEMORY_MB)) * 1024 * 1024L;
-// 0.6 is the default value used by Spark,
-// look at {@link
-//
https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/SparkConf.scala#L507}
-double memoryFraction = Double.parseDouble(
-SparkEnv.get().conf().get(SPARK_EXECUTOR_MEMORY_FRACTION_PROP,
DEFAULT_SPARK_EXECUTOR_MEMORY_FRACTION));
-double maxMemoryFractionForMerge =
Double.parseDouble(maxMemoryFraction);
-double userAvailableMemory = executorMemoryInBytes * (1 -
memoryFraction);
-long maxMemoryForMerge = (long) Math.floor(userAvailableMemory *
maxMemoryFractionForMerge);
-return Math.max(DEFAULT_MIN_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES,
maxMemoryForMerge);
- } else {
-return DEFAULT_MAX_MEMORY_FOR_SPILLABLE_MAP_IN_BYTES;
- }
+ return ConfigUtils.getMaxMemoryAllowedForMerge(props, maxMemoryFraction);
Review comment:
Okay.. if we call a Spark specific class here, then this does not achieve
the purpose right..
i.e you cannot move `ConfigUtils` to hudi-spark and keep config in
`hudi-writer-common` without making hudi-writer-common depend on ConfigUtils?
We should change the caller of `getMaxMemoryAllowedForMerge` to use
ConfigUtils.getXX()` just like how you did for storage level?
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:
us...@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-hudi] vinothchandar commented on a change in pull request #1418: [HUDI-678] Make config package spark free
vinothchandar commented on a change in pull request #1418: [HUDI-678] Make
config package spark free
URL: https://github.com/apache/incubator-hudi/pull/1418#discussion_r398216438
##
File path:
hudi-client/src/main/java/org/apache/hudi/config/HoodieWriteConfig.java
##
@@ -565,6 +556,7 @@ public FileSystemViewStorageConfig
getClientSpecifiedViewStorageConfig() {
private boolean isMemoryConfigSet = false;
private boolean isViewConfigSet = false;
private boolean isConsistencyGuardSet = false;
+private boolean isEngineConfigSet = false;
Review comment:
remove if unused?
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:
us...@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-hudi] vinothchandar commented on a change in pull request #1418: [HUDI-678] Make config package spark free
vinothchandar commented on a change in pull request #1418: [HUDI-678] Make
config package spark free
URL: https://github.com/apache/incubator-hudi/pull/1418#discussion_r398215708
##
File path:
hudi-client/src/main/java/org/apache/hudi/index/bloom/HoodieBloomIndex.java
##
@@ -69,7 +70,8 @@ public HoodieBloomIndex(HoodieWriteConfig config) {
// Step 0: cache the input record RDD
if (config.getBloomIndexUseCaching()) {
- recordRDD.persist(config.getBloomIndexInputStorageLevel());
+ StorageLevel storageLevel =
ConfigUtils.getBloomIndexInputStorageLevel(config.getProps());
Review comment:
single line? (also the usages above)
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:
us...@infra.apache.org
With regards,
Apache Git Services
[GitHub] [incubator-hudi] vinothchandar commented on a change in pull request #1418: [HUDI-678] Make config package spark free
vinothchandar commented on a change in pull request #1418: [HUDI-678] Make
config package spark free
URL: https://github.com/apache/incubator-hudi/pull/1418#discussion_r396149536
##
File path:
hudi-client/src/main/java/org/apache/hudi/client/config/AbstractConfig.java
##
@@ -0,0 +1,38 @@
+/*
+ * 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.
+ */
+
+package org.apache.hudi.client.config;
+
+import java.util.Properties;
+
+/**
+ * Abstract config for multi-engine.
+ */
+public abstract class AbstractConfig {
+ protected Properties props;
+
+ public AbstractConfig(Properties props) {
+this.props = props;
+ }
+
+ public abstract long getMaxMemoryAllowedForMerge(String maxMemoryFraction);
+
+ public abstract T getWriteStatusStorageLevel();
+
+ public abstract T getBloomIndexInputStorageLevel();
+}
Review comment:
I feel we can just make it Spark free for now.. Multi-engine can be dealt
with separately..?
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:
us...@infra.apache.org
With regards,
Apache Git Services
