Github user brkyvz commented on a diff in the pull request:
https://github.com/apache/spark/pull/9373#discussion_r43587988
--- Diff:
streaming/src/main/scala/org/apache/spark/streaming/util/FileBasedWriteAheadLog.scala
---
@@ -126,11 +127,11 @@ private[streaming] class FileBasedWriteAheadLog(
val logFilesToRead = pastLogs.map{ _.path} ++ currentLogPath
logInfo("Reading from the logs: " + logFilesToRead.mkString("\n"))
- logFilesToRead.iterator.map { file =>
+ logFilesToRead.par.map { file =>
--- End diff --
@harishreedharan I did several benchmarks with this. In a setting where you
would have a 30 minute window operation, which corresponds to ~9000 files on S3
with `closeFileAfterWrite = true`, the current code takes about 15 minutes to
recover. With this very simple addition, this time is reduced to 1.5 minutes
(on a driver with 4 CPUs, probably with hyper-threading)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
