[SPARK-49489][SQL][HIVE] HMS client respects hive.thrift.client.maxmessage.size
#50022
+80
−12
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
|
@shrprasa @wangyum @Madhukar525722 I just made it to work, would be great if you could have a test with your internal cases, I will polish the code according to the GHA result and your feedback |
|
Thanks @pan3793. We will test and update |
LuciferYang
reviewed
Feb 20, 2025
| case proxy if JdkProxy.isProxyClass(proxy.getClass) => | ||
| JdkProxy.getInvocationHandler(proxy) match { | ||
| case syncHandler if syncHandler.getClass.getName.endsWith("SynchronizedHandler") => | ||
| val realMscField = SparkClassUtils.classForName( |
There was a problem hiding this comment.
Perhaps we can directly use syncHandler.getClass().getDeclaredField("client") here, and there should be no practical difference in effect.
There was a problem hiding this comment.
Then, perhaps we can try to refactor a bit here, maybe like
JdkProxy.getInvocationHandler(proxy) match {
case syncHandler if syncHandler.getClass.getName.endsWith("SynchronizedHandler") =>
val realMsc = getFieldValue(syncHandler, "client").asInstanceOf[IMetaStoreClient]
configureMaxThriftMessageSize(hiveConf, realMsc)
case retryHandler: RetryingMetaStoreClient =>
val realMsc = getFieldValue(retryHandler, "base").asInstanceOf[IMetaStoreClient]
configureMaxThriftMessageSize(hiveConf, realMsc)
case _ => // do nothing
}
private def getFieldValue(obj: AnyRef, fieldName: String): AnyRef = {
val field = obj.getClass.getDeclaredField(fieldName)
field.setAccessible(true)
field.get(obj)
}hmm... Perhaps the exception type caught in the catch block needs to be changed.
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?
Port HIVE-26633 for Spark HMS client.
Why are the changes needed?
See discussion in #46468 (comment)
Does this PR introduce any user-facing change?
No, it fixes the bug introduced by an unreleased change, namely SPARK-47018.
How was this patch tested?
This must be tested manually, as the current Spark UT does not cover the remote HMS cases.
I constructed a test case in a testing Hadoop cluster with a remote HMS.
Firstly, create a table with a large number of partitions.
Then try to tune
hive.thrift.client.max.message.sizeand run a query that would triggergetPartitionsthrift call. For example, when set to 1mb, it throwsTTransportException: MaxMessageSize reached, and the exception disappers after boosting the value.Was this patch authored or co-authored using generative AI tooling?
No.