ORC-1430: Use Hadoop 3.3.5 shaded clients #1509
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dongjoon-hyun
changed the title
|
Could you review this, @williamhyun and @wgtmac ? |
|
Thank you so much! Merged to main for Apache ORC 2.0. |
|
Thanks @dongjoon-hyun. Good to know a better approach to use Hadoop clients. |
|
Thank you, @wgtmac ! |
dongjoon-hyun
added a commit
that referenced
this pull request
Oct 1, 2023
### What changes were proposed in this pull request? This PR aims to remove Apache Zookeeper Runtime dependency for Apache ORC 2.0.0. ### Why are the changes needed? Apache ORC 1.4.0 added Zookeeper depdency to the pom file to reduce the uber file size because Zookeeper is used by Hadoop Common module at that time. - #96 After ORC-1430, Apache ORC 2.0.0 uses Hadoop shaded clients which shaded Zookeeper runtime dependency. We can remove Zookeeper dependency. - #1509 - #1554 ### How was this patch tested? Pass the CIs. This closes #1572 . Closes #1630 from dongjoon-hyun/ORC-1514. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
dongjoon-hyun
added a commit
that referenced
this pull request
Dec 27, 2023
### What changes were proposed in this pull request? This PR aims to use `Hadoop Vectored IO` always in Apache ORC 2.0.0. ### Why are the changes needed? Apache ORC 2.0.0 is ready to use this new Hadoop feature. - #1509 - #1554 - [Hadoop Vectored IO Presentation](https://docs.google.com/presentation/d/1U5QRN4etbM7gkbnGO3OW4sCfUZx9LqJN/) > Works great everywhere; radical benefit in object stores ### How was this patch tested? Pass the CIs. Closes #1708 from williamhyun/hadoopvectorized. Lead-authored-by: William Hyun <[email protected]> Co-authored-by: Dongjoon Hyun <[email protected]> Co-authored-by: HarshitGupta11 <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
williamhyun
pushed a commit
that referenced
this pull request
Jan 4, 2024
…2_7 classes
### What changes were proposed in this pull request?
This PR aims to remove `HadoopShimsPre2_3`, `HadoopShimsPre2_6`, `HadoopShimsPre2_7` classes and use `HadoopShimsCurrent` always.
### Why are the changes needed?
1. `HadoopShimsCurrent` supports not only Apache Hadoop 3+ but also Apache Hadoop 2.7+.
2. Apache ORC 2.0 uses Hadoop 3.x shaded client doesn't need old shims for Hadoop 2.6 and olders.
- #1509
In addition, Apache Spark community also has been using the shaded Hadoop client since Spark 3.2 (SPARK-33212) and dropped `Hadoop 2` profile via [SPARK-42452](https://issues.apache.org/jira/browse/SPARK-42452) completely at Spark 3.5.0.
### How was this patch tested?
Pass the CIs.
Closes #1724 from dongjoon-hyun/ORC-1569.
Authored-by: Dongjoon Hyun <[email protected]>
Signed-off-by: William Hyun <[email protected]>
cxzl25
pushed a commit
to cxzl25/orc
that referenced
this pull request
Jan 11, 2024
### What changes were proposed in this pull request?
Currently, Apache ORC project uses three properties.
```
<hadoop.version>2.7.3</hadoop.version>
<min.hadoop.version>2.7.3</min.hadoop.version>
<tools.hadoop.version>2.7.3</tools.hadoop.version>
```
This aims the following for Apache ORC 2.0.0.
1. Use Hadoop 3.3.5 shaded clients.
2. Remove `min.hadoop.version` and `tools.hadoop.version` in favor of `hadoop.version`
3. Ban non-shaded clients from now.
```
<bannedDependencies>
<excludes>
<exclude>org.apache.hadoop:hadoop-common</exclude>
<exclude>org.apache.hadoop:hadoop-hdfs-client</exclude>
<exclude>org.apache.hadoop:hadoop-mapreduce-client-core</exclude>
<exclude>org.apache.hadoop:hadoop-mapreduce-client-jobclient</exclude>
</excludes>
<searchTransitive>true</searchTransitive>
</bannedDependencies>
```
Note that all changes are `pom.xml` files. There is no code change.
### Why are the changes needed?
- Hadoop 3's shaded client removes lots of complexity from the downstream clients.
- It's stable because Apache Spark community has been using Hadoop 3's shaded client from Apache Spark 3.2.0 (October 13, 2021) via https://issues.apache.org/jira/browse/SPARK-33212.
### How was this patch tested?
Pass the CIs.
Also, I validated there is no side-effect at Apache Spark side. The following is the change set when Apache Spark upgrades from Apache ORC 1.8.3 (AS-IS) to Apache ORC 2.0.0-SNAPSHOT.
```
-aircompressor/0.21//aircompressor-0.21.jar
+aircompressor/0.24//aircompressor-0.24.jar
-orc-core/1.8.3/shaded-protobuf/orc-core-1.8.3-shaded-protobuf.jar
-orc-mapreduce/1.8.3/shaded-protobuf/orc-mapreduce-1.8.3-shaded-protobuf.jar
-orc-shims/1.8.3//orc-shims-1.8.3.jar
+orc-core/2.0.0-SNAPSHOT/shaded-protobuf/orc-core-2.0.0-SNAPSHOT-shaded-protobuf.jar
+orc-mapreduce/2.0.0-SNAPSHOT/shaded-protobuf/orc-mapreduce-2.0.0-SNAPSHOT-shaded-protobuf.jar
+orc-shims/2.0.0-SNAPSHOT//orc-shims-2.0.0-SNAPSHOT.jar
```
Closes apache#1509 from dongjoon-hyun/ORC-1430.
Authored-by: Dongjoon Hyun <[email protected]>
Signed-off-by: Dongjoon Hyun <[email protected]>
cxzl25
pushed a commit
to cxzl25/orc
that referenced
this pull request
Jan 11, 2024
### What changes were proposed in this pull request? This PR aims to remove Apache Zookeeper Runtime dependency for Apache ORC 2.0.0. ### Why are the changes needed? Apache ORC 1.4.0 added Zookeeper depdency to the pom file to reduce the uber file size because Zookeeper is used by Hadoop Common module at that time. - apache#96 After ORC-1430, Apache ORC 2.0.0 uses Hadoop shaded clients which shaded Zookeeper runtime dependency. We can remove Zookeeper dependency. - apache#1509 - apache#1554 ### How was this patch tested? Pass the CIs. This closes apache#1572 . Closes apache#1630 from dongjoon-hyun/ORC-1514. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
cxzl25
pushed a commit
to cxzl25/orc
that referenced
this pull request
Jan 11, 2024
### What changes were proposed in this pull request? This PR aims to use `Hadoop Vectored IO` always in Apache ORC 2.0.0. ### Why are the changes needed? Apache ORC 2.0.0 is ready to use this new Hadoop feature. - apache#1509 - apache#1554 - [Hadoop Vectored IO Presentation](https://docs.google.com/presentation/d/1U5QRN4etbM7gkbnGO3OW4sCfUZx9LqJN/) > Works great everywhere; radical benefit in object stores ### How was this patch tested? Pass the CIs. Closes apache#1708 from williamhyun/hadoopvectorized. Lead-authored-by: William Hyun <[email protected]> Co-authored-by: Dongjoon Hyun <[email protected]> Co-authored-by: HarshitGupta11 <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
cxzl25
pushed a commit
to cxzl25/orc
that referenced
this pull request
Jan 11, 2024
…2_7 classes
### What changes were proposed in this pull request?
This PR aims to remove `HadoopShimsPre2_3`, `HadoopShimsPre2_6`, `HadoopShimsPre2_7` classes and use `HadoopShimsCurrent` always.
### Why are the changes needed?
1. `HadoopShimsCurrent` supports not only Apache Hadoop 3+ but also Apache Hadoop 2.7+.
2. Apache ORC 2.0 uses Hadoop 3.x shaded client doesn't need old shims for Hadoop 2.6 and olders.
- apache#1509
In addition, Apache Spark community also has been using the shaded Hadoop client since Spark 3.2 (SPARK-33212) and dropped `Hadoop 2` profile via [SPARK-42452](https://issues.apache.org/jira/browse/SPARK-42452) completely at Spark 3.5.0.
### How was this patch tested?
Pass the CIs.
Closes apache#1724 from dongjoon-hyun/ORC-1569.
Authored-by: Dongjoon Hyun <[email protected]>
Signed-off-by: William Hyun <[email protected]>
dongjoon-hyun
added a commit
to apache/spark
that referenced
this pull request
Mar 8, 2024
### What changes were proposed in this pull request? This PR aims to Upgrade Apache ORC to 2.0.0 for Apache Spark 4.0.0. Apache ORC community has 3-year support policy which is longer than Apache Spark. It's aligned like the following. - Apache ORC 2.0.x <-> Apache Spark 4.0.x - Apache ORC 1.9.x <-> Apache Spark 3.5.x - Apache ORC 1.8.x <-> Apache Spark 3.4.x - Apache ORC 1.7.x (Supported) <-> Apache Spark 3.3.x (End-Of-Support) ### Why are the changes needed? **Release Note** - https://github.com/apache/orc/releases/tag/v2.0.0 **Milestone** - https://github.com/apache/orc/milestone/20?closed=1 - apache/orc#1728 - apache/orc#1801 - apache/orc#1498 - apache/orc#1627 - apache/orc#1497 - apache/orc#1509 - apache/orc#1554 - apache/orc#1708 - apache/orc#1733 - apache/orc#1760 - apache/orc#1743 ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Pass the CIs. ### Was this patch authored or co-authored using generative AI tooling? No. Closes #45443 from dongjoon-hyun/SPARK-44115. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
sweisdb
pushed a commit
to sweisdb/spark
that referenced
this pull request
Apr 1, 2024
### What changes were proposed in this pull request? This PR aims to Upgrade Apache ORC to 2.0.0 for Apache Spark 4.0.0. Apache ORC community has 3-year support policy which is longer than Apache Spark. It's aligned like the following. - Apache ORC 2.0.x <-> Apache Spark 4.0.x - Apache ORC 1.9.x <-> Apache Spark 3.5.x - Apache ORC 1.8.x <-> Apache Spark 3.4.x - Apache ORC 1.7.x (Supported) <-> Apache Spark 3.3.x (End-Of-Support) ### Why are the changes needed? **Release Note** - https://github.com/apache/orc/releases/tag/v2.0.0 **Milestone** - https://github.com/apache/orc/milestone/20?closed=1 - apache/orc#1728 - apache/orc#1801 - apache/orc#1498 - apache/orc#1627 - apache/orc#1497 - apache/orc#1509 - apache/orc#1554 - apache/orc#1708 - apache/orc#1733 - apache/orc#1760 - apache/orc#1743 ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Pass the CIs. ### Was this patch authored or co-authored using generative AI tooling? No. Closes apache#45443 from dongjoon-hyun/SPARK-44115. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
Currently, Apache ORC project uses three properties.
This aims the following for Apache ORC 2.0.0.
min.hadoop.versionandtools.hadoop.versionin favor ofhadoop.versionNote that all changes are
pom.xmlfiles. There is no code change.Why are the changes needed?
How was this patch tested?
Pass the CIs.
Also, I validated there is no side-effect at Apache Spark side. The following is the change set when Apache Spark upgrades from Apache ORC 1.8.3 (AS-IS) to Apache ORC 2.0.0-SNAPSHOT.