From hdfs-dev-return-33134-archive-asf-public=cust-asf.ponee.io@hadoop.apache.org Thu Jul 5 12:13:04 2018 Return-Path: X-Original-To: archive-asf-public@cust-asf.ponee.io Delivered-To: archive-asf-public@cust-asf.ponee.io Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by mx-eu-01.ponee.io (Postfix) with SMTP id 01688180657 for ; Thu, 5 Jul 2018 12:13:03 +0200 (CEST) Received: (qmail 43770 invoked by uid 500); 5 Jul 2018 10:13:02 -0000 Mailing-List: contact hdfs-dev-help@hadoop.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Delivered-To: mailing list hdfs-dev@hadoop.apache.org Received: (qmail 43759 invoked by uid 99); 5 Jul 2018 10:13:02 -0000 Received: from pnap-us-west-generic-nat.apache.org (HELO spamd3-us-west.apache.org) (209.188.14.142) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 05 Jul 2018 10:13:02 +0000 Received: from localhost (localhost [127.0.0.1]) by spamd3-us-west.apache.org (ASF Mail Server at spamd3-us-west.apache.org) with ESMTP id B86EA18094D for ; Thu, 5 Jul 2018 10:13:01 +0000 (UTC) X-Virus-Scanned: Debian amavisd-new at spamd3-us-west.apache.org X-Spam-Flag: NO X-Spam-Score: -110.311 X-Spam-Level: X-Spam-Status: No, score=-110.311 tagged_above=-999 required=6.31 tests=[ENV_AND_HDR_SPF_MATCH=-0.5, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_SPF_WL=-7.5, USER_IN_WHITELIST=-100] autolearn=disabled Received: from mx1-lw-us.apache.org ([10.40.0.8]) by localhost (spamd3-us-west.apache.org [10.40.0.10]) (amavisd-new, port 10024) with ESMTP id mcRwudAFAQQ7 for ; Thu, 5 Jul 2018 10:13:01 +0000 (UTC) Received: from mailrelay1-us-west.apache.org (mailrelay1-us-west.apache.org [209.188.14.139]) by mx1-lw-us.apache.org (ASF Mail Server at mx1-lw-us.apache.org) with ESMTP id 6E9765F4A3 for ; Thu, 5 Jul 2018 10:13:01 +0000 (UTC) Received: from jira-lw-us.apache.org (unknown [207.244.88.139]) by mailrelay1-us-west.apache.org (ASF Mail Server at mailrelay1-us-west.apache.org) with ESMTP id DF550E1095 for ; Thu, 5 Jul 2018 10:13:00 +0000 (UTC) Received: from jira-lw-us.apache.org (localhost [127.0.0.1]) by jira-lw-us.apache.org (ASF Mail Server at jira-lw-us.apache.org) with ESMTP id 24E1D27508 for ; Thu, 5 Jul 2018 10:13:00 +0000 (UTC) Date: Thu, 5 Jul 2018 10:13:00 +0000 (UTC) From: "Hari Sekhon (JIRA)" To: hdfs-dev@hadoop.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Created] (HDFS-13720) HDFS dataset Anti-Affinity Block Placement across DataNodes for data local task optimization (improve Spark executor utilization & performance) MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 Hari Sekhon created HDFS-13720: ---------------------------------- Summary: HDFS dataset Anti-Affinity Block Placement across Dat= aNodes for data local task optimization (improve Spark executor utilization= & performance) Key: HDFS-13720 URL: https://issues.apache.org/jira/browse/HDFS-13720 Project: Hadoop HDFS Issue Type: Improvement Components: balancer & mover, block placement, performance Affects Versions: 2.7.3 Environment: Hortonworks HDP 2.6 Reporter: Hari Sekhon Improvement Request for Anti-Affinity Block Placement across datanodes such= that for a given data set the blocks are distributed evenly across all ava= ilable datanodes in order to improve task scheduling while maintaining data= locality. This could be done via a=C2=A0client side write flag=C2=A0as well as=C2=A0v= ia a balancer command switch combined with giving a target path to files or= directories to=C2=A0redistributed as evenly as possible across all datanod= es in the cluster. See this following Spark issue which causes massive under-utilisation acros= s jobs. Only 30-50% of executor cores were being used for tasks due to data= locality targeting. Many executors doing literally nothing, while holding = significant cluster resources, because the=C2=A0data set, which in at least= one job was large enough to have 30,000 tasks churning though=C2=A0slowly = on only a subset of the available executors. The workaround in the end was = to disable data local tasks in Spark, but if everyone did that the bottlene= ck would go back to being the network and it undermines Hadoop's=C2=A0first= premise of don't move the data to compute. For performance critical jobs, = returning tasks to Yarn isn't a good idea either, they want the jobs to use= all the resources available, not just the resources on a subset of nodes t= hat hold a given dataset or pulling half the blocks across the network. https://issues.apache.org/jira/browse/SPARK-24474 -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-dev-unsubscribe@hadoop.apache.org For additional commands, e-mail: hdfs-dev-help@hadoop.apache.org