Skip to content

allow non-roundrobin manner of mesos scheduling #24

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

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

allow non-roundrobin manner of mesos scheduling #24

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
99 changes: 65 additions & 34 deletions ...n/scala/org/apache/spark/scheduler/cluster/mesos/MesosCoarseGrainedSchedulerBackend.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -107,6 +107,13 @@ private[spark] class MesosCoarseGrainedSchedulerBackend(
private val slaveOfferConstraints =
parseConstraintString(sc.conf.get("spark.mesos.constraints", ""))

// Sort offers by CPUs in a descending manner.
// This allows tasks to be assigned to most used workers first.
private val sortOffersDescending = conf.getBoolean("spark.mesos.sortOffersDescending", false)

// Assign tasks in round robin manner. If set false, tasks will be assigned sequentially.
private val assignTasksRoundRobin = conf.getBoolean("spark.mesos.assignTaskRoundRobin", true)

// Reject offers with mismatched constraints in seconds
private val rejectOfferDurationForUnmetConstraints =
getRejectOfferDurationForUnmetConstraints(sc)
Expand Down Expand Up @@ -324,8 +331,14 @@ private[spark] class MesosCoarseGrainedSchedulerBackend(
*/
private def handleMatchedOffers(
d: org.apache.mesos.SchedulerDriver, offers: mutable.Buffer[Offer]): Unit = {
val tasks = buildMesosTasks(offers)
for (offer <- offers) {
val offersMaybeSorted = if (sortOffersDescending) {
offers.sortWith { (l, r) =>
getResource(l.getResourcesList, "cpus") > getResource(r.getResourcesList, "cpus")
}
} else offers
val tasks = buildMesosTasks(offersMaybeSorted)

for (offer <- offersMaybeSorted) {
val offerAttributes = toAttributeMap(offer.getAttributesList)
val offerMem = getResource(offer.getResourcesList, "mem")
val offerCpus = getResource(offer.getResourcesList, "cpus")
Expand Down Expand Up @@ -388,46 +401,64 @@ private[spark] class MesosCoarseGrainedSchedulerBackend(
for (offer <- offers) {
val slaveId = offer.getSlaveId.getValue
val offerId = offer.getId.getValue
val resources = remainingResources(offerId)

if (canLaunchTask(slaveId, resources)) {
// Create a task
launchTasks = true
val taskId = newMesosTaskId()
val offerCPUs = getResource(resources, "cpus").toInt
var resources = remainingResources(offerId)

// sometimes you don't want to round-robin create
Copy link

Choose a reason for hiding this comment

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

The method comment probably becomes ambiguous:

  • Returns a map from OfferIDs to the tasks to launch on those offers. In order to maximize
  • per-task memory and IO, tasks are round-robin assigned to offers.

if (assignTasksRoundRobin) {
if (canLaunchTask(slaveId, resources)) {
launchTasks = true
// Create a task
createTask(offer, tasks, resources, remainingResources)
}
} else {
while (canLaunchTask(slaveId, resources)) {
launchTasks = true
createTask(offer, tasks, resources, remainingResources)
resources = remainingResources(offerId)
}
}
}
}
tasks.toMap
}

val taskCPUs = executorCores(offerCPUs)
val taskMemory = executorMemory(sc)
private def createTask(
offer: Offer,
tasks: mutable.Map[OfferID, List[MesosTaskInfo]],
resources: JList[Resource],
remainingResources: mutable.Map[String, JList[Resource]]): Unit = {
val slaveId = offer.getSlaveId.getValue
val taskId = newMesosTaskId()
val offerCPUs = getResource(resources, "cpus").toInt

slaves.getOrElseUpdate(slaveId, new Slave(offer.getHostname)).taskIDs.add(taskId)
val taskCPUs = executorCores(offerCPUs)
val taskMemory = executorMemory(sc)

val (resourcesLeft, resourcesToUse) =
partitionTaskResources(resources, taskCPUs, taskMemory)
slaves.getOrElseUpdate(slaveId, new Slave(offer.getHostname)).taskIDs.add(taskId)

val taskBuilder = MesosTaskInfo.newBuilder()
.setTaskId(TaskID.newBuilder().setValue(taskId.toString).build())
.setSlaveId(offer.getSlaveId)
.setCommand(createCommand(offer, taskCPUs + extraCoresPerExecutor, taskId))
.setName("Task " + taskId)
val (resourcesLeft, resourcesToUse) =
partitionTaskResources(resources, taskCPUs, taskMemory)

taskBuilder.addAllResources(resourcesToUse.asJava)
val taskBuilder = MesosTaskInfo.newBuilder()
.setTaskId(TaskID.newBuilder().setValue(taskId.toString).build())
.setSlaveId(offer.getSlaveId)
.setCommand(createCommand(offer, taskCPUs + extraCoresPerExecutor, taskId))
.setName("Task " + taskId)

sc.conf.getOption("spark.mesos.executor.docker.image").foreach { image =>
MesosSchedulerBackendUtil.setupContainerBuilderDockerInfo(
image,
sc.conf,
taskBuilder.getContainerBuilder
)
}
taskBuilder.addAllResources(resourcesToUse.asJava)

tasks(offer.getId) ::= taskBuilder.build()
remainingResources(offerId) = resourcesLeft.asJava
totalCoresAcquired += taskCPUs
coresByTaskId(taskId) = taskCPUs
}
}
sc.conf.getOption("spark.mesos.executor.docker.image").foreach { image =>
MesosSchedulerBackendUtil.setupContainerBuilderDockerInfo(
image,
sc.conf,
taskBuilder.getContainerBuilder
)
}
tasks.toMap

tasks(offer.getId) ::= taskBuilder.build()
remainingResources(offer.getId.getValue) = resourcesLeft.asJava
totalCoresAcquired += taskCPUs
coresByTaskId(taskId) = taskCPUs
}

/** Extracts task needed resources from a list of available resources. */
Expand Down
18 changes: 18 additions & 0 deletions ...la/org/apache/spark/scheduler/cluster/mesos/MesosCoarseGrainedSchedulerBackendSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -176,6 +176,24 @@ class MesosCoarseGrainedSchedulerBackendSuite extends SparkFunSuite
verifyTaskLaunched(driver, "o2")
}

test("mesos does not assign tasks round-robin on offers") {
val executorCores = 4
val maxCores = executorCores * 2
setBackend(Map("spark.executor.cores" -> executorCores.toString,
"spark.cores.max" -> maxCores.toString,
"spark.mesos.assignTaskRoundRobin" -> "false",
"spark.mesos.sortOffersDescending" -> "false"
))

val executorMemory = backend.executorMemory(sc)
offerResources(List(
(executorMemory * 2, executorCores * 2),
(executorMemory * 2, executorCores * 2)))

verifyTaskLaunched(driver, "o1")
verifyTaskLaunched(driver, "o1")
}

test("mesos creates multiple executors on a single slave") {
val executorCores = 4
setBackend(Map("spark.executor.cores" -> executorCores.toString))
Expand Down