Update Spark to version 3.1.1 #247
Merged
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
This skips version 3.0.2 and goes directly to 3.1.1. There are no compilation issues and the apache-spark benchmarks execute normally.
This was referenced Apr 30, 2021
Closed
farquet
approved these changes
Apr 30, 2021
|
Do you have any idea what's wrong with the |
|
A bit of investigation (but no final solution yet) in #249 |
SBT uses 1024 by default and kills memory-related JVM options in the `.sbtopts` file and derives its values on its own. The default memory size results in code cache of 128 megs, which is insufficient for a pristine build, where it can reach slightly over 200 megs. Raising the memory to 2048 increases the code cache size to 256 megs.
This will avoid having to change the version number in multiple places. The current version number is determine by the latest (annotated) tag and the distance to it from the current commit. Release built from the tagged commit will have a "clean" version number, others will include the hash of the build commit. CI scripts will be able to determine the current version simply by using `git describe` and stripping the initial `v`. The only thing that needs to be updated is the git.baseVersion setting in renaissance-core/version.sbt, which provides a base (upcoming) version number for git-less situations, i.e., when building the sources outside the git tree.
This will avoid having to update README.md with each commit.
Removes task to write out .gitprops which does not work in git-less mode and is not needed anyway. Also adds a proper version related to the upstream master and the exported dependency.
These checks are also made in the CI, but it does not hurt to have them in the pre-push hook. The markdown check requires a successful build of the base JAR.
The goal is to use the `script` section of `.travis.yml` for higher-level actions instead of low-level file management.
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.
This skips version 3.0.2 and goes directly to 3.1.1. There are no compilation issues and the apache-spark benchmarks execute normally. It might (but I did not check) with JDK compatibility issues. The attached CSV files show the times of the first iteration on OpenJDK 1.8.0_282 (just to show that it does not crash).
spark301.csv
spark311.csv
I think that until we freeze for a release, we should stay up-to-date whenever possible.