[CDAP-20922] Reuse the Spark ClassLoader in app-fabric #15483
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
Contributor
chtyim
commented
Dec 4, 2023
chtyim
commented
- For non local spark program execution purposes, we don’t need to isolate the Spark classloader, since there is no creation of the SparkContext object.
dli357
reviewed
Dec 4, 2023
| // Never needs to rewrite yarn client in CDAP master, which is the only place using | ||
| // distributed program runner. Also wrap the classloader so that it can't be closed. | ||
| distributedRunnerClassLoader = new ClassLoader(createClassLoader(true, false, false)) {}; | ||
| return distributedRunnerClassLoader; |
There was a problem hiding this comment.
Instead of returning distributedRunnerClassLoader and incurring another expensive read from a volatile variable, can we write the new classloader to our existing local variable classLoader, assign it to distributedRunnerClassLoader, then return the local variable?
There was a problem hiding this comment.
That's not really expensive and it is just have it hit once due to the singleton. Just don't wanna clutter the code too much
chtyim
force-pushed
the
feature/cache-spark-cl
branch
2 times, most recently
from
9ddb843 to
a4ecc5d
Compare
- For non local spark program execution purposes, we don’t need to isolate the Spark classloader, since there is no creation of the SparkContext object.
chtyim
force-pushed
the
feature/cache-spark-cl
branch
from
a4ecc5d to
7e9abd0
Compare
arjan-bal
reviewed
Dec 5, 2023
| } | ||
| try { | ||
| // Never needs to rewrite yarn client in CDAP master, which is the only place using | ||
| // distributed program runner. Also wrap the classloader so that it can't be closed. |
There was a problem hiding this comment.
Wrapping it will also make it difficult for task workers to close the class loader. So we won't be able to solve the memory increases in task workers until we have a way to re-use this class loader in task workers also.
arjan-bal
changed the title
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.