Github user kayousterhout commented on a diff in the pull request:
https://github.com/apache/spark/pull/14079#discussion_r92081806
--- Diff:
core/src/test/scala/org/apache/spark/scheduler/TaskSchedulerImplSuite.scala ---
@@ -408,6 +411,96 @@ class TaskSchedulerImplSuite extends SparkFunSuite
with LocalSparkContext with B
}
assert(tsm.isZombie)
}
+
+ // the tasksSets complete, so the tracker should be notified of the
successful ones
+ verify(blacklist, times(1)).updateBlacklistForSuccessfulTaskSet(
+ stageId = 0,
+ stageAttemptId = 0,
+ failuresByExec = stageToMockTaskSetBlacklist(0).execToFailures)
+ verify(blacklist, times(1)).updateBlacklistForSuccessfulTaskSet(
+ stageId = 1,
+ stageAttemptId = 0,
+ failuresByExec = stageToMockTaskSetBlacklist(1).execToFailures)
+ // but we shouldn't update for the failed taskset
+ verify(blacklist, never).updateBlacklistForSuccessfulTaskSet(
+ stageId = meq(2),
+ stageAttemptId = anyInt(),
+ failuresByExec = anyObject())
+ }
+
+ test("scheduled tasks obey node and executor blacklists") {
--- End diff --
why not have this test be a normal case (i.e., when the tsm doesn't get
aborted, but the blacklist is obeyed)? That seems like the more common case,
and abortion is tested in the test below
---
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
