spark-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
Subject spark git commit: [SPARK-16956] Make ApplicationState.MAX_NUM_RETRY configurable
Date Tue, 09 Aug 2016 18:22:16 GMT
Repository: spark
Updated Branches:
  refs/heads/master 92da22878 -> b89b3a5c8

[SPARK-16956] Make ApplicationState.MAX_NUM_RETRY configurable

## What changes were proposed in this pull request?

This patch introduces a new configuration, `spark.deploy.maxExecutorRetries`, to let users
configure an obscure behavior in the standalone master where the master will kill Spark applications
which have experienced too many back-to-back executor failures. The current setting is a hardcoded
constant (10); this patch replaces that with a new cluster-wide configuration.

**Background:** This application-killing was added in 6b5980da796e0204a7735a31fb454f312bc9daac
(from September 2012) and I believe that it was designed to prevent a faulty application whose
executors could never launch from DOS'ing the Spark cluster via an infinite series of executor
launch attempts. In a subsequent patch (#1360), this feature was refined to prevent applications
which have running executors from being killed by this code path.

**Motivation for making this configurable:** Previously, if a Spark Standalone application
experienced more than `ApplicationState.MAX_NUM_RETRY` executor failures and was left with
no executors running then the Spark master would kill that application, but this behavior
is problematic in environments where the Spark executors run on unstable infrastructure and
can all simultaneously die. For instance, if your Spark driver runs on an on-demand EC2 instance
while all workers run on ephemeral spot instances then it's possible for all executors to
die at the same time while the driver stays alive. In this case, it may be desirable to keep
the Spark application alive so that it can recover once new workers and executors are available.
In order to accommodate this use-case, this patch modifies the Master to never kill faulty
applications if `spark.deploy.maxExecutorRetries` is negative.

I'd like to merge this patch into master, branch-2.0, and branch-1.6.

## How was this patch tested?

I tested this manually using `spark-shell` and `local-cluster` mode. This is a tricky feature
to unit test and historically this code has not changed very often, so I'd prefer to skip
the additional effort of adding a testing framework and would rather rely on manual tests
and review for now.

Author: Josh Rosen <>

Closes #14544 from JoshRosen/add-setting-for-max-executor-failures.


Branch: refs/heads/master
Commit: b89b3a5c8e391fcaebe7ef3c77ef16bb9431d6ab
Parents: 92da228
Author: Josh Rosen <>
Authored: Tue Aug 9 11:21:45 2016 -0700
Committer: Josh Rosen <>
Committed: Tue Aug 9 11:21:45 2016 -0700

 .../spark/deploy/master/ApplicationState.scala       |  2 --
 .../org/apache/spark/deploy/master/Master.scala      |  7 ++++++-
 docs/                             | 15 +++++++++++++++
 3 files changed, 21 insertions(+), 3 deletions(-)
diff --git a/core/src/main/scala/org/apache/spark/deploy/master/ApplicationState.scala b/core/src/main/scala/org/apache/spark/deploy/master/ApplicationState.scala
index 37bfcdf..097728c 100644
--- a/core/src/main/scala/org/apache/spark/deploy/master/ApplicationState.scala
+++ b/core/src/main/scala/org/apache/spark/deploy/master/ApplicationState.scala
@@ -22,6 +22,4 @@ private[master] object ApplicationState extends Enumeration {
   type ApplicationState = Value
-  val MAX_NUM_RETRY = 10
diff --git a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala
index fded847..dfffc47 100644
--- a/core/src/main/scala/org/apache/spark/deploy/master/Master.scala
+++ b/core/src/main/scala/org/apache/spark/deploy/master/Master.scala
@@ -58,6 +58,7 @@ private[deploy] class Master(
   private val RETAINED_DRIVERS = conf.getInt("spark.deploy.retainedDrivers", 200)
   private val REAPER_ITERATIONS = conf.getInt("spark.dead.worker.persistence", 15)
   private val RECOVERY_MODE = conf.get("spark.deploy.recoveryMode", "NONE")
+  private val MAX_EXECUTOR_RETRIES = conf.getInt("spark.deploy.maxExecutorRetries", 10)
   val workers = new HashSet[WorkerInfo]
   val idToApp = new HashMap[String, ApplicationInfo]
@@ -265,7 +266,11 @@ private[deploy] class Master(
             val normalExit = exitStatus == Some(0)
             // Only retry certain number of times so we don't go into an infinite loop.
-            if (!normalExit && appInfo.incrementRetryCount() >= ApplicationState.MAX_NUM_RETRY)
+            // Important note: this code path is not exercised by tests, so be very careful
+            // changing this `if` condition.
+            if (!normalExit
+                && appInfo.incrementRetryCount() >= MAX_EXECUTOR_RETRIES
+                && MAX_EXECUTOR_RETRIES >= 0) { // < 0 disables this application-killing
               val execs = appInfo.executors.values
               if (!execs.exists(_.state == ExecutorState.RUNNING)) {
                 logError(s"Application ${} with ID ${} failed
" +
diff --git a/docs/ b/docs/
index c864c90..5ae63fe 100644
--- a/docs/
+++ b/docs/
@@ -196,6 +196,21 @@ SPARK_MASTER_OPTS supports the following system properties:
+  <td><code>spark.deploy.maxExecutorRetries</code></td>
+  <td>10</td>
+  <td>
+    Limit on the maximum number of back-to-back executor failures that can occur before the
+    standalone cluster manager removes a faulty application. An application will never be
+    if it has any running executors. If an application experiences more than
+    <code>spark.deploy.maxExecutorRetries</code> failures in a row, no executors
+    successfully start running in between those failures, and the application has no running
+    executors then the standalone cluster manager will remove the application and mark it
as failed.
+    To disable this automatic removal, set <code>spark.deploy.maxExecutorRetries</code>
+    <code>-1</code>.
+    <br/>
+  </td>

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message