Mailing-List: contact issues-help@spark.apache.org; run by ezmlm
Precedence: bulk
Date: Fri, 1 Dec 2017 02:56:00 +0000 (UTC)
From: "Wenchen Fan (JIRA)" <jira@apache.org>
To: issues@spark.apache.org
Message-ID: <JIRA.13121678.1511970672000.361841.1512096960950@Atlassian.JIRA>
In-Reply-To: <JIRA.13121678.1511970672000@Atlassian.JIRA>
References: <JIRA.13121678.1511970672000@Atlassian.JIRA> <JIRA.13121678.1511970672781@jira-lw-us.apache.org>
Subject: [jira] [Assigned] (SPARK-22653) executorAddress registered in
 CoarseGrainedSchedulerBackend.executorDataMap is null
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Fri, 01 Dec 2017 02:56:06 -0000


     [ https://issues.apache.org/jira/browse/SPARK-22653?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Wenchen Fan reassigned SPARK-22653:
-----------------------------------

    Assignee: Thomas Graves

> executorAddress registered in CoarseGrainedSchedulerBackend.executorDataMap is null
> -----------------------------------------------------------------------------------
>
>                 Key: SPARK-22653
>                 URL: https://issues.apache.org/jira/browse/SPARK-22653
>             Project: Spark
>          Issue Type: Bug
>          Components: Scheduler
>    Affects Versions: 2.2.0
>            Reporter: Thomas Graves
>            Assignee: Thomas Graves
>             Fix For: 2.3.0
>
>
> In CoarseGrainedSchedulerBackend.RegisterExecutor the executor data address (executorRef.address) can be null.
>  val data = new ExecutorData(executorRef, executorRef.address, hostname,
>             cores, cores, logUrls)
> At this point the executorRef.address can be null, there is actually code above it that handles this case:
>  // If the executor's rpc env is not listening for incoming connections, `hostPort`
>           // will be null, and the client connection should be used to contact the executor.
>           val executorAddress = if (executorRef.address != null) {
>               executorRef.address
>             } else {
>               context.senderAddress
>             }
> But it doesn't use executorAddress when it creates the ExecutorData.
> This causes removeExecutor to never remove it properly from addressToExecutorId.
>             addressToExecutorId -= executorInfo.executorAddress
> This is also a memory leak and can also call onDisconnected to call disableExecutor when it shouldn't.


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org