[
https://issues.apache.org/jira/browse/SPARK-15170?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Owen resolved SPARK-15170.
-------------------------------
Resolution: Not A Problem
This looks correct. It's reporting the threshold that the idle time has
crossed, not the idle time (which is presumably, a little bit larger, but
doesn't matter much). That info isn't available at this point in the code
anyway.
> Log error message in ExecutorAllocationManager
> ----------------------------------------------
>
> Key: SPARK-15170
> URL: https://issues.apache.org/jira/browse/SPARK-15170
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
> Reporter: meiyoula
> Priority: Trivial
>
> No matter how long the expire idle time of executor, the log just says "it
> has been idle for $executorIdleTimeoutS seconds". Because
> executorIdleTimeoutS =
> conf.getTimeAsSeconds("spark.dynamicAllocation.executorIdleTimeout", "60s"),
> so it logs same expire time for different executor.
> {quote}
> logInfo(s"Removing executor $executorId because it has been idle for " +
> s"$executorIdleTimeoutS seconds (new desired total will be
> ${numExistingExecutors - 1})")
> {quote}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
