[
https://issues.apache.org/jira/browse/HADOOP-13000?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15224695#comment-15224695
]
Steve Loughran commented on HADOOP-13000:
-----------------------------------------
This is code trying to talk to an RM that's down, with
{{yarn.resourcemanager.connect.max-wait.ms == 0 }} and
{{yarn.resourcemanager.connect.retry-interval.ms==0}}
I'd have expected a fail-fast with the underlying cause, instead I see an
{{UndeclaredThrowableException}}
{code}
2016-04-04 18:58:55,265 [YarnHistoryProvider Refresher] WARN
server.YarnHistoryProvider (Logging.scala:logWarning(91)) - Exception in
refresh thread
java.lang.reflect.UndeclaredThrowableException
at com.sun.proxy.$Proxy19.getApplications(Unknown Source)
at
org.apache.hadoop.yarn.client.api.impl.YarnClientImpl.getApplications(YarnClientImpl.java:479)
at
org.apache.hadoop.yarn.client.api.impl.YarnClientImpl.getApplications(YarnClientImpl.java:463)
at
org.apache.spark.deploy.history.yarn.server.YarnHistoryProvider.listYarnSparkApplications(YarnHistoryProvider.scala:947)
at
org.apache.spark.deploy.history.yarn.server.YarnHistoryProvider$$anonfun$listAndCacheApplications$1.apply(YarnHistoryProvider.scala:633)
at
org.apache.spark.deploy.history.yarn.server.YarnHistoryProvider$$anonfun$listAndCacheApplications$1.apply(YarnHistoryProvider.scala:617)
at
org.apache.spark.deploy.history.yarn.ExtendedMetricsSource$class.time(ExtendedMetricsSource.scala:60)
at
org.apache.spark.deploy.history.yarn.server.YarnHistoryProviderMetrics.time(YarnHistoryProvider.scala:1264)
at
org.apache.spark.deploy.history.yarn.server.YarnHistoryProvider.listAndCacheApplications(YarnHistoryProvider.scala:616)
at
org.apache.spark.deploy.history.yarn.server.YarnHistoryProvider$Refresher.doRefresh(YarnHistoryProvider.scala:1185)
at
org.apache.spark.deploy.history.yarn.server.YarnHistoryProvider$Refresher.run(YarnHistoryProvider.scala:1131)
at java.lang.Thread.run(Thread.java:745)
Caused by: java.lang.InterruptedException: sleep interrupted
at java.lang.Thread.sleep(Native Method)
at
org.apache.hadoop.io.retry.RetryInvocationHandler.invoke(RetryInvocationHandler.java:159)
... 12 more
{code}
> IPC client can raise UndeclaredThrowableException during failure handling if
> maxwait==0
> ---------------------------------------------------------------------------------------
>
> Key: HADOOP-13000
> URL: https://issues.apache.org/jira/browse/HADOOP-13000
> Project: Hadoop Common
> Issue Type: Bug
> Components: ipc
> Affects Versions: 2.7.1
> Reporter: Steve Loughran
>
> When an IPC falls, it may goes into a retry process, which sleeps via
> Thread.sleep() .
> If this happens, and the max wait interval is set to 0, then an
> InterruptedException can trigger in a Thread.sleep() operation. This isn't in
> the method signatures, so gets wrapped in an UndeclaredThrowable. This isn't
> an IOE, so all code catching IOEs get confused.
> The original exception is also completely lost.
> Proposed: InterruptedExceptions in the sleep are caught, trigger throwing of
> the original exception instead
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
