HeartSaVioR commented on a change in pull request #25797: [SPARK-29043][Core]
Improve the concurrent performance of History Server
URL: https://github.com/apache/spark/pull/25797#discussion_r324490443
##########
File path:
core/src/main/scala/org/apache/spark/deploy/history/FsHistoryProvider.scala
##########
@@ -155,6 +161,16 @@ private[history] class FsHistoryProvider(conf: SparkConf,
clock: Clock)
Utils.deleteRecursively(dbPath)
open(dbPath, metadata)
}
+ }
+
+ // Visible for testing.
+ private[history] val listing: KVStore = storePath.map { path =>
+ getKVStore(path, "listing", CURRENT_LISTING_VERSION)
+ }.getOrElse(new InMemoryStore())
+
+ private[history] val processing: KVStore = storePath.map { path =>
Review comment:
Given it's only viable when SHS is running (no need to restore between
restarting SHS), it doesn't really need to be stored into KVStore - in-memory
would be OK. (Maybe a Map for logPath -> `Future` instance received from
replayExecutor, so that we can deal with graceful shutdown if necessary.) You
can reduce many changes needed for this.
----------------------------------------------------------------
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]
