From mapreduce-issues-return-93826-archive-asf-public=cust-asf.ponee.io@hadoop.apache.org Tue Oct 29 20:28:03 2019 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 [207.244.88.153]) by mx-eu-01.ponee.io (Postfix) with SMTP id E31991806C4 for ; Tue, 29 Oct 2019 21:28:02 +0100 (CET) Received: (qmail 49423 invoked by uid 500); 29 Oct 2019 20:28:02 -0000 Mailing-List: contact mapreduce-issues-help@hadoop.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Delivered-To: mailing list mapreduce-issues@hadoop.apache.org Received: (qmail 49400 invoked by uid 99); 29 Oct 2019 20:28:02 -0000 Received: from mailrelay1-us-west.apache.org (HELO mailrelay1-us-west.apache.org) (209.188.14.139) by apache.org (qpsmtpd/0.29) with ESMTP; Tue, 29 Oct 2019 20:28:02 +0000 Received: from jira-he-de.apache.org (static.172.67.40.188.clients.your-server.de [188.40.67.172]) by mailrelay1-us-west.apache.org (ASF Mail Server at mailrelay1-us-west.apache.org) with ESMTP id 2D6F6E300F for ; Tue, 29 Oct 2019 20:28:00 +0000 (UTC) Received: from jira-he-de.apache.org (localhost.localdomain [127.0.0.1]) by jira-he-de.apache.org (ASF Mail Server at jira-he-de.apache.org) with ESMTP id 35DD8780538 for ; Tue, 29 Oct 2019 20:28:00 +0000 (UTC) Date: Tue, 29 Oct 2019 20:28:00 +0000 (UTC) From: "Ahmed Hussein (Jira)" To: mapreduce-issues@hadoop.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (MAPREDUCE-7208) Tuning TaskRuntimeEstimator MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/MAPREDUCE-7208?page=3Dcom.atlas= sian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D= 16962437#comment-16962437 ]=20 Ahmed Hussein commented on MAPREDUCE-7208: ------------------------------------------ Thanks [~jeagles]. I looked at the test cases: * {{hadoop.mapreduce.v2.TestSpeculativeExecutionWithMRApp}} is a related te= st case and It was failing because I changed the threshold of the estimate = that triggers a new speculative task. I fixed that default behavior in the = new patch. * {{hadoop.mapred.TestLocalMRNotification}} and {{hadoop.mapreduce.v2.TestM= ROldApiJobs}} seem to be a random failure. They pass successfully on local = machine. > Tuning TaskRuntimeEstimator=20 > ---------------------------- > > Key: MAPREDUCE-7208 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-7208 > Project: Hadoop Map/Reduce > Issue Type: Improvement > Reporter: Ahmed Hussein > Assignee: Ahmed Hussein > Priority: Minor > Attachments: MAPREDUCE-7208.001.patch, MAPREDUCE-7208.002.patch, = smoothing-exponential.md > > > By default, MR uses LegacyTaskRuntimeEstimator to get an estimate of the = runtime. =C2=A0The estimator does not adjust dynamically to the progress ra= te of the tasks. On the other hand, the existing alternative "Exponentially= SmoothedTaskRuntimeEstimator" behavior in unpredictable. > =C2=A0 > There are several dimensions to improve the exponential implementation: > # Exponential shooting needs a warmup period. Otherwise, the estimate wi= ll be affected by the initial values. > # Using a single=C2=A0smoothing factor (Lambda) does not work well for a= ll the tasks. To increase the level of smoothing across the majority of tas= ks, we need to give a range of flexibility to dynamically adjust the smooth= ing factor=C2=A0based on the history of the task progress. > # Design wise, it is better to separate between the statistical model an= d the MR interface.=C2=A0We need to have a way to evaluate estimators stati= stically, without the need to run MR. For example, an estimator can be eval= uated as a black box by using a stream of raw data as input and testing the= accuracy of the generated stream of estimates. > # The exponential estimator=C2=A0speculates frequently and fails to dete= ct slowing tasks. It=C2=A0does not detect slowing tasks. As a result, a tas= kAttempt that does not do any progress won't trigger a new speculation. > =C2=A0 > The file [^smoothing-exponential.md]=C2=A0describes how Simple Exponentia= l smoothing factor works. > =C2=A0 > =C2=A0 -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: mapreduce-issues-unsubscribe@hadoop.apache.org For additional commands, e-mail: mapreduce-issues-help@hadoop.apache.org