Igor Berman created SPARK-23423:
-----------------------------------
Summary: Application declines any offers when killed+active
executors rich spark.dynamicAllocation.maxExecutors
Key: SPARK-23423
URL: https://issues.apache.org/jira/browse/SPARK-23423
Project: Spark
Issue Type: Bug
Components: Mesos, Spark Core
Affects Versions: 2.2.1
Reporter: Igor Berman
Hi
I've noticed rather strange behavior of MesosCoarseGrainedSchedulerBackend when
running on Mesos with dynamic allocation on and limiting number of max
executors by spark.dynamicAllocation.maxExecutors.
Suppose we have long running driver that has cyclic pattern of resource
consumption(with some idle times in between), due to dyn.allocation it receives
offers and then releases them after current chunk of work processed.
Since at
[https://github.com/apache/spark/blob/master/resource-managers/mesos/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosCoarseGrainedSchedulerBackend.scala#L573]
the backend compares numExecutors < executorLimit and
numExecutors is defined as slaves.values.map(_.taskIDs.size).sum and slaves
holds all executors ever "met", i.e. both active and killed (see comment
[https://github.com/apache/spark/blob/master/resource-managers/mesos/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosCoarseGrainedSchedulerBackend.scala#L122)]
it means that after a while, when number of killed executors might be greater
than maxExecutors, the application will decline any offer, thus stopping to work
I've created test that "reproduces" this behavior, not sure how good it is:
{code:java}
//MesosCoarseGrainedSchedulerBackendSuite
test("max executors registered stops to accept offers when dynamic allocation
enabled") {
setBackend(Map(
"spark.dynamicAllocation.maxExecutors" -> "1",
"spark.dynamicAllocation.enabled" -> "true",
"spark.dynamicAllocation.testing" -> "true"))
backend.doRequestTotalExecutors(1)
val (mem, cpu) = (backend.executorMemory(sc), 4)
val offer1 = createOffer("o1", "s1", mem, cpu)
backend.resourceOffers(driver, List(offer1).asJava)
verifyTaskLaunched(driver, "o1")
backend.doKillExecutors(List("0"))
verify(driver, times(1)).killTask(createTaskId("0"))
val offer2 = createOffer("o2", "s2", mem, cpu)
backend.resourceOffers(driver, List(offer2).asJava)
verify(driver, times(1)).declineOffer(offer2.getId)
}{code}
Workaround: Don't set maxExecutors with dynamicAllocation on
I'm not sure how to solve this problem since it seems that it's not trivial to
change numExecutors in this scenario to count only active executors(since this
information is not available in Slave class. On the other hand, might be that
this behavior is "normal" and expected.
Please advice
Igor
marking you friends since you were last to touch this piece of code and
probably can advice something([~vanzin], [~skonto], [~susanxhuynh])
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
