From dev-return-321377-archive-asf-public=cust-asf.ponee.io@lucene.apache.org Mon May 7 14:36:05 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 0B8AB180648 for ; Mon, 7 May 2018 14:36:04 +0200 (CEST) Received: (qmail 12349 invoked by uid 500); 7 May 2018 12:36:03 -0000 Mailing-List: contact dev-help@lucene.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@lucene.apache.org Delivered-To: mailing list dev@lucene.apache.org Received: (qmail 12339 invoked by uid 99); 7 May 2018 12:36:03 -0000 Received: from pnap-us-west-generic-nat.apache.org (HELO spamd1-us-west.apache.org) (209.188.14.142) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 07 May 2018 12:36:03 +0000 Received: from localhost (localhost [127.0.0.1]) by spamd1-us-west.apache.org (ASF Mail Server at spamd1-us-west.apache.org) with ESMTP id 23020C6F98 for ; Mon, 7 May 2018 12:36:03 +0000 (UTC) X-Virus-Scanned: Debian amavisd-new at spamd1-us-west.apache.org X-Spam-Flag: NO X-Spam-Score: -109.501 X-Spam-Level: X-Spam-Status: No, score=-109.501 tagged_above=-999 required=6.31 tests=[ENV_AND_HDR_SPF_MATCH=-0.5, KAM_ASCII_DIVIDERS=0.8, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, USER_IN_DEF_SPF_WL=-7.5, USER_IN_WHITELIST=-100] autolearn=disabled Received: from mx1-lw-eu.apache.org ([10.40.0.8]) by localhost (spamd1-us-west.apache.org [10.40.0.7]) (amavisd-new, port 10024) with ESMTP id cobaUDZBG0Fp for ; Mon, 7 May 2018 12:36:02 +0000 (UTC) Received: from mailrelay1-us-west.apache.org (mailrelay1-us-west.apache.org [209.188.14.139]) by mx1-lw-eu.apache.org (ASF Mail Server at mx1-lw-eu.apache.org) with ESMTP id 52AC15F3BC for ; Mon, 7 May 2018 12:36:01 +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 75E8DE1209 for ; Mon, 7 May 2018 12:36:00 +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 2F1A721297 for ; Mon, 7 May 2018 12:36:00 +0000 (UTC) Date: Mon, 7 May 2018 12:36:00 +0000 (UTC) From: "Simon Willnauer (JIRA)" To: dev@lucene.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (LUCENE-7976) Make TieredMergePolicy respect maxSegmentSizeMB and allow singleton merges of very large segments 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/LUCENE-7976?page=3Dcom.atlassia= n.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D164= 65852#comment-16465852 ]=20 Simon Willnauer commented on LUCENE-7976: ----------------------------------------- >=C2=A0[~erickerickson]=C2=A0if you index with a single thread, and `commit= ()` at the right times you can build a precise set of segments and then dir= ectly test TMP's behavior.=C2=A0 I like approach one since it then gives yo= u full deterministic control to enumerate the different tricky cases that s= urface in real indices? =C2=A0 I really think we should start working towards testing this as real unittes= t. Creating stuff with IW and depending on it is a big issue. We can change= the code to be less dependent on IW. I think we should and we should do it= before making significant changes to MPs IMO > Make TieredMergePolicy respect maxSegmentSizeMB and allow singleton merge= s of very large segments > -------------------------------------------------------------------------= ------------------------ > > Key: LUCENE-7976 > URL: https://issues.apache.org/jira/browse/LUCENE-7976 > Project: Lucene - Core > Issue Type: Improvement > Reporter: Erick Erickson > Assignee: Erick Erickson > Priority: Major > Attachments: LUCENE-7976.patch, LUCENE-7976.patch, LUCENE-7976.pa= tch, LUCENE-7976.patch, LUCENE-7976.patch, LUCENE-7976.patch, LUCENE-7976.p= atch > > > We're seeing situations "in the wild" where there are very large indexes = (on disk) handled quite easily in a single Lucene index. This is particular= ly true as features like docValues move data into MMapDirectory space. The = current TMP algorithm allows on the order of 50% deleted documents as per a= dev list conversation with Mike McCandless (and his blog here: https://ww= w.elastic.co/blog/lucenes-handling-of-deleted-documents). > Especially in the current era of very large indexes in aggregate, (think = many TB) solutions like "you need to distribute your collection over more s= hards" become very costly. Additionally, the tempting "optimize" button exa= cerbates the issue since once you form, say, a 100G segment (by optimizing/= forceMerging) it is not eligible for merging until 97.5G of the docs in it = are deleted (current default 5G max segment size). > The proposal here would be to add a new parameter to TMP, something like = (no, that's not serious name, suggestio= ns welcome) which would default to 100 (or the same behavior we have now). > So if I set this parameter to, say, 20%, and the max segment size stays a= t 5G, the following would happen when segments were selected for merging: > > any segment with > 20% deleted documents would be merged or rewritten N= O MATTER HOW LARGE. There are two cases, > >> the segment has < 5G "live" docs. In that case it would be merged with= smaller segments to bring the resulting segment up to 5G. If no smaller se= gments exist, it would just be rewritten > >> The segment has > 5G "live" docs (the result of a forceMerge or optimi= ze). It would be rewritten into a single segment removing all deleted docs = no matter how big it is to start. The 100G example above would be rewritten= to an 80G segment for instance. > Of course this would lead to potentially much more I/O which is why the d= efault would be the same behavior we see now. As it stands now, though, the= re's no way to recover from an optimize/forceMerge except to re-index from = scratch. We routinely see 200G-300G Lucene indexes at this point "in the wi= ld" with 10s of shards replicated 3 or more times. And that doesn't even i= nclude having these over HDFS. > Alternatives welcome! Something like the above seems minimally invasive. = A new merge policy is certainly an alternative. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org For additional commands, e-mail: dev-help@lucene.apache.org