Github user andrewor14 commented on a diff in the pull request:
https://github.com/apache/spark/pull/3861#discussion_r24644419
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/CoarseMesosSchedulerBackend.scala
---
@@ -193,45 +252,107 @@ private[spark] class CoarseMesosSchedulerBackend(
override def reregistered(d: SchedulerDriver, masterInfo: MasterInfo) {}
+ lazy val shuffleServiceEnabled =
conf.getBoolean("spark.shuffle.service.enabled", false)
+
+ lazy val shuffleServiceMem = if (shuffleServiceEnabled) {
+ conf.getInt("spark.mesos.shuffle.service.mem", 1024)
+ } else {
+ 0
+ }
+
+ lazy val shuffleServiceCpu = if (shuffleServiceEnabled) {
+ conf.getInt("spark.mesos.shuffle.service.cpu", 1)
+ } else {
+ 0
+ }
+
/**
* Method called by Mesos to offer resources on slaves. We respond by
launching an executor,
* unless we've already launched more than we wanted to.
*/
override def resourceOffers(d: SchedulerDriver, offers: JList[Offer]) {
- synchronized {
+ stateLock.synchronized {
val filters = Filters.newBuilder().setRefuseSeconds(-1).build()
for (offer <- offers) {
- val slaveId = offer.getSlaveId.toString
+ val slaveId = offer.getSlaveId.getValue
val mem = getResource(offer.getResourcesList, "mem")
val cpus = getResource(offer.getResourcesList, "cpus").toInt
- if (totalCoresAcquired < maxCores &&
- mem >= MemoryUtils.calculateTotalMemory(sc) &&
- cpus >= 1 &&
- failuresBySlaveId.getOrElse(slaveId, 0) < MAX_SLAVE_FAILURES &&
- !slaveIdsWithExecutors.contains(slaveId)) {
+ logTrace("Received offer id: " + offer.getId.getValue + ", cpu: "
+ cpus.toString +
+ ", mem: " + mem.toString)
+
+ var minCpuRequired = 1
+ var minMemRequired = MemoryUtils.calculateTotalMemory(sc)
+ var shuffleResourcesUsed = false
+
+ if (!slaveStatuses.contains(slaveId) ||
!slaveStatuses(slaveId).executorRunning) {
+ minCpuRequired += shuffleServiceCpu
+ minMemRequired += shuffleServiceMem
+ shuffleResourcesUsed = true
+ }
+
+ if (taskIdToSlaveId.size < executorLimit.getOrElse(Int.MaxValue) &&
+ totalCoresAcquired < maxCores &&
+ failuresBySlaveId.getOrElse(slaveId, 0) < MAX_SLAVE_FAILURES &&
+ (!slaveStatuses.contains(slaveId) ||
!slaveStatuses(slaveId).taskRunning) &&
+ mem >= minMemRequired &&
+ cpus >= minCpuRequired) {
--- End diff --
this if case is very hard to read. Put a few of these cases in properly
named variables and chain them together?
---
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]
