dbtsai commented on a change in pull request #26235: [SPARK-29576][Core] Use
Spark's CompressionCodec for Ser/Deser of MapOutputStatus
URL: https://github.com/apache/spark/pull/26235#discussion_r338291945
##########
File path: core/src/main/scala/org/apache/spark/MapOutputTracker.scala
##########
@@ -952,11 +936,12 @@ private[spark] object MapOutputTracker extends Logging {
}
// Opposite of serializeMapStatuses.
- def deserializeMapStatuses(bytes: Array[Byte]): Array[MapStatus] = {
+ def deserializeMapStatuses(bytes: Array[Byte], conf: SparkConf):
Array[MapStatus] = {
assert (bytes.length > 0)
def deserializeObject(arr: Array[Byte], off: Int, len: Int): AnyRef = {
- val objIn = new ObjectInputStream(new ZstdInputStream(
+ val codec = CompressionCodec.createCodec(conf, "zstd")
Review comment:
Yes. The codec is wrapped in buffered stream which avoids overhead excessive
of JNI call while trying to compress small amount of data.
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
