Return-Path: X-Original-To: apmail-giraph-dev-archive@www.apache.org Delivered-To: apmail-giraph-dev-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id F39D4180E2 for ; Fri, 11 Dec 2015 06:19:36 +0000 (UTC) Received: (qmail 14818 invoked by uid 500); 11 Dec 2015 06:19:36 -0000 Delivered-To: apmail-giraph-dev-archive@giraph.apache.org Received: (qmail 14766 invoked by uid 500); 11 Dec 2015 06:19:36 -0000 Mailing-List: contact dev-help@giraph.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@giraph.apache.org Delivered-To: mailing list dev@giraph.apache.org Received: (qmail 14754 invoked by uid 99); 11 Dec 2015 06:19:36 -0000 Received: from Unknown (HELO spamd3-us-west.apache.org) (209.188.14.142) by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 11 Dec 2015 06:19:36 +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 EE3B2180B60 for ; Fri, 11 Dec 2015 06:19:35 +0000 (UTC) X-Virus-Scanned: Debian amavisd-new at spamd3-us-west.apache.org X-Spam-Flag: NO X-Spam-Score: 2.879 X-Spam-Level: ** X-Spam-Status: No, score=2.879 tagged_above=-999 required=6.31 tests=[DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=3, RCVD_IN_MSPIKE_H3=-0.01, RCVD_IN_MSPIKE_WL=-0.01, SPF_PASS=-0.001] autolearn=disabled Authentication-Results: spamd3-us-west.apache.org (amavisd-new); dkim=pass (2048-bit key) header.d=gmail.com Received: from mx1-eu-west.apache.org ([10.40.0.8]) by localhost (spamd3-us-west.apache.org [10.40.0.10]) (amavisd-new, port 10024) with ESMTP id 41F_x-XWY8ob for ; Fri, 11 Dec 2015 06:19:34 +0000 (UTC) Received: from mail-oi0-f53.google.com (mail-oi0-f53.google.com [209.85.218.53]) by mx1-eu-west.apache.org (ASF Mail Server at mx1-eu-west.apache.org) with ESMTPS id 08F3F24E9F for ; Fri, 11 Dec 2015 06:19:34 +0000 (UTC) Received: by oiww189 with SMTP id w189so59282942oiw.3 for ; Thu, 10 Dec 2015 22:19:26 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :content-type; bh=N8N+n+btuvn0X+qvEiHo7eJhGRFP7bPkBdgjuiZLKek=; b=ms2lSablj7ibQhFrz82cbXXttOzXX6xXmYN+HexrfRpfS/2RAYKqgq+1DbZL1CazH+ rKgIbVDCY45ndaR/9jDNyqlkDsQa5owF9AVaA/XBnVEp2fcdOZfocGczySKat9b9ogOY e2lO3U3SQ5fIo2ecb98CXeBPBaeX6IXcnNwfPlTutpHj5bP6xtvwh01nNSjXqbnVoOgs JRoZwrKrUdWLj37H+4jYP+IIxf0awiH7L5oy2stkByx1eVmIa0XoJUrZw7/EO4TWjMr4 32nVqM0khlY1Mp9K2ApddqiFpKhjAnVxNYp34diftHEr/o7L6UZ6Vb0lzR1F1bbrKAGo B+8w== X-Received: by 10.202.195.202 with SMTP id t193mr12490122oif.94.1449814766446; Thu, 10 Dec 2015 22:19:26 -0800 (PST) MIME-Version: 1.0 Received: by 10.202.188.10 with HTTP; Thu, 10 Dec 2015 22:18:47 -0800 (PST) In-Reply-To: References: From: Hassan Eslami Date: Fri, 11 Dec 2015 00:18:47 -0600 Message-ID: Subject: Re: OOC on Giraph 1.1 or 1.2 To: dev@giraph.apache.org Content-Type: multipart/alternative; boundary=001a1137c360250b510526994e8f --001a1137c360250b510526994e8f Content-Type: text/plain; charset=UTF-8 Yes, OOC mechanism in 1.1 may not work in certain scenarios: - If a worker goes into long GC pauses (which is likely since a job that is supposed to use OOC feature usually suffer from high memory pressure), all incoming messages will be buffered before coming to Giraph runtime, and this will cause he job to crash. - The rate at which a worker receives messages may be more than the rate at which a worker spills messages to disk (specially in applications where compute function does a little work, but generates a fair number of messages, such as PageRank). In this case, messages keep accumulating and the worker crashes. - OOC feature is not supported for input superstep for Giraph 1.0 and 1.1. That being said, if graph size is larger than aggregate available memory, you are likely to see the job crashes during input superstep. OOC feature is under active development at this moment. A new OOC mechanism is recently pushed to the code base (which basically solves the problem for input superstep and OOC graph). However, the new OOC mechanism still have problems for OOC messages. An entirely new approach to OOC is being developed and soon will be available in the code base. Best, Hassan On Thu, Dec 10, 2015 at 8:58 PM, Khaled Ammar wrote: > Hi, > > I remember that I used OOC on Giraph 1.0, how ever it is not working on > 1.1. > > Is there a known bug in the OOC feature for recent Giraph versions? > > -- > Thanks, > -Khaled > --001a1137c360250b510526994e8f--