Skip to content

Commit

Permalink
Add different severity levels
Browse files Browse the repository at this point in the history
With this commit, the HistoryNotFoundPage behaves differently for
each of the following scenarios:
  (1) User did not enable event logging -> point them to the configs
  (2) Event logs are not found for some reason -> tell them that
  (3) Exception is thrown in replaying -> show them the stack trace
  • Loading branch information
andrewor14 committed Jul 17, 2014
1 parent 832b687 commit 97cddc0
Show file tree
Hide file tree
Showing 4 changed files with 91 additions and 33 deletions.
7 changes: 2 additions & 5 deletions core/src/main/scala/org/apache/spark/TaskEndReason.scala
Original file line number Diff line number Diff line change
Expand Up @@ -20,6 +20,7 @@ package org.apache.spark
import org.apache.spark.annotation.DeveloperApi
import org.apache.spark.executor.TaskMetrics
import org.apache.spark.storage.BlockManagerId
import org.apache.spark.util.Utils

/**
* :: DeveloperApi ::
Expand Down Expand Up @@ -88,11 +89,7 @@ case class ExceptionFailure(
stackTrace: Array[StackTraceElement],
metrics: Option[TaskMetrics])
extends TaskFailedReason {
override def toErrorString: String = {
val stackTraceString =
if (stackTrace == null) "null" else stackTrace.map(" " + _).mkString("\n")
s"$className ($description)\n$stackTraceString"
}
override def toErrorString: String = Utils.exceptionString(className, description, stackTrace)
}

/**
Expand Down
61 changes: 38 additions & 23 deletions core/src/main/scala/org/apache/spark/deploy/master/Master.scala
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@

package org.apache.spark.deploy.master

import java.net.URLEncoder
import java.text.SimpleDateFormat
import java.util.Date

Expand All @@ -30,7 +31,6 @@ import akka.actor._
import akka.pattern.ask
import akka.remote.{DisassociatedEvent, RemotingLifecycleEvent}
import akka.serialization.SerializationExtension
import org.apache.hadoop.fs.FileSystem

import org.apache.spark.{Logging, SecurityManager, SparkConf, SparkException}
import org.apache.spark.deploy.{ApplicationDescription, DriverDescription, ExecutorState}
Expand Down Expand Up @@ -642,10 +642,7 @@ private[spark] class Master(
waitingApps -= app

// If application events are logged, use them to rebuild the UI
if (!rebuildSparkUI(app)) {
// Avoid broken links if the UI is not reconstructed
app.desc.appUiUrl = "/history/not-found"
}
rebuildSparkUI(app)

for (exec <- app.executors.values) {
exec.worker.removeExecutor(exec)
Expand All @@ -667,29 +664,47 @@ private[spark] class Master(
*/
def rebuildSparkUI(app: ApplicationInfo): Boolean = {
val appName = app.desc.name
val eventLogDir = app.desc.eventLogDir.getOrElse { return false }
val eventLogDir = app.desc.eventLogDir.getOrElse {
// Event logging is not enabled for this application
app.desc.appUiUrl = "/history/not-found"
return false
}
val fileSystem = Utils.getHadoopFileSystem(eventLogDir)
val eventLogInfo = EventLoggingListener.parseLoggingInfo(eventLogDir, fileSystem)
val eventLogPaths = eventLogInfo.logPaths
val compressionCodec = eventLogInfo.compressionCodec
if (!eventLogPaths.isEmpty) {
try {
val replayBus = new ReplayListenerBus(eventLogPaths, fileSystem, compressionCodec)
val ui = new SparkUI(
new SparkConf, replayBus, appName + " (completed)", "/history/" + app.id)
replayBus.replay()
app.desc.appUiUrl = ui.basePath
appIdToUI(app.id) = ui
webUi.attachSparkUI(ui)
return true
} catch {
case e: Exception =>
logError("Exception in replaying log for application %s (%s)".format(appName, app.id), e)
}
} else {
logWarning("Application %s (%s) has no valid logs: %s".format(appName, app.id, eventLogDir))

if (eventLogPaths.isEmpty) {
// Event logging is enabled for this application, but no event logs are found
val title = s"Application history not found (${app.id})"
var msg = s"No event logs found for application $appName in $eventLogDir."
logWarning(msg)
msg += " Did you specify the correct logging directory?"
msg = URLEncoder.encode(msg, "UTF-8")
app.desc.appUiUrl = s"/history/not-found?msg=$msg&title=$title"
return false
}

try {
val replayBus = new ReplayListenerBus(eventLogPaths, fileSystem, compressionCodec)
val ui = new SparkUI(new SparkConf, replayBus, appName + " (completed)", "/history/" + app.id)
replayBus.replay()
appIdToUI(app.id) = ui
webUi.attachSparkUI(ui)
// Application UI is successfully rebuilt, so link the Master UI to it
app.desc.appUiUrl = ui.basePath
true
} catch {
case e: Exception =>
// Relay exception message to application UI page
val title = s"Application history load error (${app.id})"
val exception = URLEncoder.encode(Utils.exceptionString(e), "UTF-8")
var msg = s"Exception in replaying log for application $appName!"
logError(msg, e)
msg = URLEncoder.encode(msg, "UTF-8")
app.desc.appUiUrl = s"/history/not-found?msg=$msg&exception=$exception&title=$title"
false
}
false
}

/** Generate a new app ID given a app's submission date */
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@

package org.apache.spark.deploy.master.ui

import java.net.URLDecoder
import javax.servlet.http.HttpServletRequest

import scala.xml.Node
Expand All @@ -26,17 +27,47 @@ import org.apache.spark.ui.{UIUtils, WebUIPage}
private[spark] class HistoryNotFoundPage(parent: MasterWebUI)
extends WebUIPage("history/not-found") {

/**
* Render a page that conveys failure in loading application history.
*
* This accepts 3 HTTP parameters:
* msg = message to display to the user
* title = title of the page
* exception = detailed description of the exception in loading application history (if any)
*
* Parameters "msg" and "exception" are assumed to be UTF-8 encoded.
*/
def render(request: HttpServletRequest): Seq[Node] = {
val content =
val titleParam = request.getParameter("title")
val msgParam = request.getParameter("msg")
val exceptionParam = request.getParameter("exception")

// If no parameters are specified, assume the user did not enable event logging
val defaultTitle = "Event logging is not enabled"
val defaultContent =
<div class="row-fluid">
<div class="span12" style="font-size:14px">
No event logs were found for this application. To
No event logs were found for this application! To
<a href="http://spark.apache.org/docs/latest/monitoring.html">enable event logging</a>,
please set <span style="font-style:italic">spark.eventLog.enabled</span> to true and
set <span style="font-style:italic">spark.eventLog.enabled</span> to true and
<span style="font-style:italic">spark.eventLog.dir</span> to the directory to which your
event logs are written.
</div>
</div>
UIUtils.basicSparkPage(content, "Application history not found")

val title = Option(titleParam).getOrElse(defaultTitle)
val content = Option(msgParam)
.map { msg => URLDecoder.decode(msg, "UTF-8") }
.map { msg =>
<div class="row-fluid">
<div class="span12" style="font-size:14px">{msg}</div>
</div> ++
Option(exceptionParam)
.map { e => URLDecoder.decode(e, "UTF-8") }
.map { e => <pre>{e}</pre> }
.getOrElse(Seq.empty)
}.getOrElse(defaultContent)

UIUtils.basicSparkPage(content, title)
}
}
17 changes: 16 additions & 1 deletion core/src/main/scala/org/apache/spark/util/Utils.scala
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,7 @@ import org.apache.spark.executor.ExecutorUncaughtExceptionHandler
import org.apache.spark.serializer.{DeserializationStream, SerializationStream, SerializerInstance}

/** CallSite represents a place in user code. It can have a short and a long form. */
private[spark] case class CallSite(val short: String, val long: String)
private[spark] case class CallSite(short: String, long: String)

/**
* Various utility methods used by Spark.
Expand Down Expand Up @@ -1291,4 +1291,19 @@ private[spark] object Utils extends Logging {
}
}

/** Return a nice string representation of the exception, including the stack trace. */
def exceptionString(e: Exception): String = {
if (e == null) "" else exceptionString(getFormattedClassName(e), e.getMessage, e.getStackTrace)
}

/** Return a nice string representation of the exception, including the stack trace. */
def exceptionString(
className: String,
description: String,
stackTrace: Array[StackTraceElement]): String = {
val desc = if (description == null) "" else description
val st = if (stackTrace == null) "" else stackTrace.map(" " + _).mkString("\n")
s"$className: $desc\n$st"
}

}

0 comments on commit 97cddc0

Please sign in to comment.