[GitHub] [spark] gaborgsomogyi commented on a change in pull request #31495: [SPARK-34383][SS] Optimize WAL commit phase via reducing cost of filesystem operations

[GitBox <gi...@apache.org>]

gaborgsomogyi commented on a change in pull request #31495:
URL: https://github.com/apache/spark/pull/31495#discussion_r595183474



##########
File path: sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/OffsetSeqLog.scala
##########
@@ -46,6 +47,23 @@ import org.apache.spark.sql.connector.read.streaming.{Offset => OffsetV2}
 class OffsetSeqLog(sparkSession: SparkSession, path: String)
   extends HDFSMetadataLog[OffsetSeq](sparkSession, path) {
 
+  private val cachedMetadata = new ju.TreeMap[Long, OffsetSeq]()
+
+  override def add(batchId: Long, metadata: OffsetSeq): Boolean = {
+    val added = super.add(batchId, metadata)
+    if (added) {
+      // cache metadata as it will be read again
+      cachedMetadata.put(batchId, metadata)
+      // we don't access metadata for (batchId - 2) batches; evict them

Review comment:
       I've gone through this thread and changing the current solution would introduce quite some non-trivial code structures and the gain would be not significant from memory point of view so +1 to keep it as-is.




----------------------------------------------------------------
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:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org