Return-Path: X-Original-To: apmail-cassandra-commits-archive@www.apache.org Delivered-To: apmail-cassandra-commits-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 099E419A49 for ; Tue, 29 Mar 2016 15:35:26 +0000 (UTC) Received: (qmail 20167 invoked by uid 500); 29 Mar 2016 15:35:25 -0000 Delivered-To: apmail-cassandra-commits-archive@cassandra.apache.org Received: (qmail 20136 invoked by uid 500); 29 Mar 2016 15:35:25 -0000 Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@cassandra.apache.org Delivered-To: mailing list commits@cassandra.apache.org Received: (qmail 20105 invoked by uid 99); 29 Mar 2016 15:35:25 -0000 Received: from arcas.apache.org (HELO arcas) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Tue, 29 Mar 2016 15:35:25 +0000 Received: from arcas.apache.org (localhost [127.0.0.1]) by arcas (Postfix) with ESMTP id 815692C1F64 for ; Tue, 29 Mar 2016 15:35:25 +0000 (UTC) Date: Tue, 29 Mar 2016 15:35:25 +0000 (UTC) From: "DOAN DuyHai (JIRA)" To: commits@cassandra.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (CASSANDRA-11383) Avoid index segment stitching in RAM which lead to OOM on big SSTable files 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/CASSANDRA-11383?page=3Dcom.atla= ssian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId= =3D15216185#comment-15216185 ]=20 DOAN DuyHai commented on CASSANDRA-11383: ----------------------------------------- 1. {{SPARSE}} index (in the sense for 1 index value, there are very few mat= ching pk) is working well. During indexing process, if there are more than = 5 partition keys for the same index value, SASI will throw an exception and= skip indexing the current value to move on the next value 2. mode {{PREFIX}} is working fine for {{DENSE}} numerical index (period_en= d_month_int). The index supports equality and range queries > Avoid index segment stitching in RAM which lead to OOM on big SSTable fil= es=20 > -------------------------------------------------------------------------= --- > > Key: CASSANDRA-11383 > URL: https://issues.apache.org/jira/browse/CASSANDRA-1138= 3 > Project: Cassandra > Issue Type: Bug > Components: CQL > Environment: C* 3.4 > Reporter: DOAN DuyHai > Assignee: Jordan West > Labels: sasi > Fix For: 3.5 > > Attachments: CASSANDRA-11383.patch, SASI_Index_build_LCS_1G_Max_S= STable_Size_logs.tar.gz, new_system_log_CMS_8GB_OOM.log, system.log_sasi_bu= ild_oom > > > 13 bare metal machines > - 6 cores CPU (12 HT) > - 64Gb RAM > - 4 SSD in RAID0 > JVM settings: > - G1 GC > - Xms32G, Xmx32G > Data set: > - =E2=89=88 100Gb/per node > - 1.3 Tb cluster-wide > - =E2=89=88 20Gb for all SASI indices > C* settings: > - concurrent_compactors: 1 > - compaction_throughput_mb_per_sec: 256 > - memtable_heap_space_in_mb: 2048 > - memtable_offheap_space_in_mb: 2048 > I created 9 SASI indices > - 8 indices with text field, NonTokenizingAnalyser, PREFIX mode, case-i= nsensitive > - 1 index with numeric field, SPARSE mode > After a while, the nodes just gone OOM. > I attach log files. You can see a lot of GC happening while index segmen= ts are flush to disk. At some point the node OOM ... > /cc [~xedin] -- This message was sent by Atlassian JIRA (v6.3.4#6332)