Github user tdas commented on a diff in the pull request:
https://github.com/apache/spark/pull/4964#discussion_r26527326
--- Diff:
streaming/src/main/scala/org/apache/spark/streaming/scheduler/JobGenerator.scala
---
@@ -43,6 +47,9 @@ class JobGenerator(jobScheduler: JobScheduler) extends
Logging {
private val ssc = jobScheduler.ssc
private val conf = ssc.conf
private val graph = ssc.graph
+ private lazy val completedBothCheckpoints = new HashMap[Time, Boolean]()
--- End diff --
I dont like the fact that the data structure encodes that there will be two
checkpoints for each batch. Checkpointing is something that should be doable
any time, and any number of times. Restricting a data strcuture to a constraint
that it can only be done twice is bad design.
In fact I say that my earlier design of automatically deleting checkpoints
whenever checkpoint is called, was wrong as well. That's why adding the flag in
my PR is important, so that we dissociate the act of taking a checkpoint from
the act of clearing checkpoints.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
