Skip to content

[SPARK-12122][STREAMING] Prevent batches from being submitted twice after recovering StreamingContext from checkpoint #10127

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
tdas wants to merge 2 commits into from
Closed

[SPARK-12122][STREAMING] Prevent batches from being submitted twice after recovering StreamingContext from checkpoint #10127

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d904b25
Remove duplicate
tdas Dec 3, 2015
fe69fbf
Changed filter
tdas Dec 3, 2015
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion streaming/src/main/scala/org/apache/spark/streaming/scheduler/JobGenerator.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -220,7 +220,8 @@ class JobGenerator(jobScheduler: JobScheduler) extends Logging {
logInfo("Batches pending processing (" + pendingTimes.size + " batches): " +
pendingTimes.mkString(", "))
// Reschedule jobs for these times
val timesToReschedule = (pendingTimes ++ downTimes).distinct.sorted(Time.ordering)
val timesToReschedule = (pendingTimes ++ downTimes).filter { _ < restartTime }
.distinct.sorted(Time.ordering)
Copy link
Member

@zsxwing zsxwing Dec 3, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you clarify why pendingTimes may contain restartTime?

Copy link
Contributor Author

@tdas tdas Dec 3, 2015

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Explained offline:

The restart time is always checkpointTime+1 (assuming batch duration = 1). However, pending times can already have batches >= checkpointTime+1. This can cause timesToReschedule to have batches >= checkpointTime+1, which will be explicitly submitted, and then resubmitted through the timer.

logInfo("Batches to reschedule (" + timesToReschedule.size + " batches): " +
timesToReschedule.mkString(", "))
timesToReschedule.foreach { time =>
Expand Down