Skip to content

[SPARK-7831][Mesos] Added flag to shutdown driver when mesos dispatch… #10701

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

Closed
nraychaudhuri wants to merge 2 commits into from
Closed

[SPARK-7831][Mesos] Added flag to shutdown driver when mesos dispatch… #10701

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
2cc0022
[SPARK-7831][Mesos] Added flag to shutdown driver when mesos dispatch…
nraychaudhuri Jan 11, 2016
9002258
Setting the driver failover timeout
nraychaudhuri Jan 15, 2016
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
2 changes: 1 addition & 1 deletion core/src/main/scala/org/apache/spark/deploy/mesos/MesosClusterDispatcher.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@ private[mesos] class MesosClusterDispatcher(
case _ => throw new IllegalArgumentException("Unsupported recovery mode: " + recoveryMode)
}

private val scheduler = new MesosClusterScheduler(engineFactory, conf)
private val scheduler = new MesosClusterScheduler(engineFactory, conf, args.driverFailOver)

private val server = new MesosRestServer(args.host, args.port, conf, scheduler)
private val webUi = new MesosClusterUI(
Expand Down
6 changes: 6 additions & 0 deletions core/src/main/scala/org/apache/spark/deploy/mesos/MesosClusterDispatcherArguments.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,7 @@ private[mesos] class MesosClusterDispatcherArguments(args: Array[String], conf:
var masterUrl: String = _
var zookeeperUrl: Option[String] = None
var propertiesFile: String = _
var driverFailOver = true // keep the driver registered with Mesos after dispatcher is stopped

parse(args.toList)

Expand Down Expand Up @@ -70,6 +71,10 @@ private[mesos] class MesosClusterDispatcherArguments(args: Array[String], conf:
propertiesFile = value
parse(tail)

case ("--disable-failover") :: tail =>
driverFailOver = false
parse(tail)

case ("--help") :: tail =>
printUsageAndExit(0)

Expand Down Expand Up @@ -97,6 +102,7 @@ private[mesos] class MesosClusterDispatcherArguments(args: Array[String], conf:
" --webui-port WEBUI_PORT WebUI Port to listen on (default: 8081)\n" +
" --name NAME Framework name to show in Mesos UI\n" +
" -m --master MASTER URI for connecting to Mesos master\n" +
" --disable-failover Will kill the driver when dispatcher is stopped\n" +
Copy link
Contributor

@tnachen tnachen Jan 14, 2016

Choose a reason for hiding this comment

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

Killing the driver doesn't sound right to me, how about "De-register the framework when dispatcher is stopped"

" -z --zk ZOOKEEPER Comma delimited URLs for connecting to \n" +
" Zookeeper for persistence\n" +
" --properties-file FILE Path to a custom Spark properties file.\n" +
Expand Down
5 changes: 3 additions & 2 deletions core/src/main/scala/org/apache/spark/scheduler/cluster/mesos/MesosClusterScheduler.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,7 +114,8 @@ private[spark] class MesosDriverState(
*/
private[spark] class MesosClusterScheduler(
engineFactory: MesosClusterPersistenceEngineFactory,
conf: SparkConf)
conf: SparkConf,
driverFailOver: Boolean = true)
extends Scheduler with MesosSchedulerUtils {
var frameworkUrl: String = _
private val metricsSystem =
Expand Down Expand Up @@ -318,7 +319,7 @@ private[spark] class MesosClusterScheduler(
ready = false
metricsSystem.report()
metricsSystem.stop()
mesosDriver.stop(true)
mesosDriver.stop(driverFailOver)
Copy link

@skyluc skyluc Jan 14, 2016

Choose a reason for hiding this comment

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

Instead of killing without failover, we could also start it without failover.

In the start method, to use:

val driver = createSchedulerDriver(
       master,
       MesosClusterScheduler.this,
       Utils.getCurrentUserName(),
       appName,
       conf,
       Some(frameworkUrl),
       Some(driverFailOver),                                 // <-- with or without checkpoint data
       Some(if (driverFailOver) Double.MaxValue else 0.0),   // <-- timeout for failover recovery
       fwId)

Copy link
Contributor Author

@nraychaudhuri nraychaudhuri Jan 14, 2016

Choose a reason for hiding this comment

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

Great find @skyluc

I will make the change

}

override def registered(
Expand Down