Return-Path: X-Original-To: apmail-accumulo-notifications-archive@minotaur.apache.org Delivered-To: apmail-accumulo-notifications-archive@minotaur.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id A040117D52 for ; Tue, 14 Apr 2015 14:48:45 +0000 (UTC) Received: (qmail 76144 invoked by uid 500); 14 Apr 2015 14:48:12 -0000 Delivered-To: apmail-accumulo-notifications-archive@accumulo.apache.org Received: (qmail 76106 invoked by uid 500); 14 Apr 2015 14:48:12 -0000 Mailing-List: contact notifications-help@accumulo.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: jira@apache.org Delivered-To: mailing list notifications@accumulo.apache.org Received: (qmail 76095 invoked by uid 99); 14 Apr 2015 14:48:12 -0000 Received: from arcas.apache.org (HELO arcas.apache.org) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Tue, 14 Apr 2015 14:48:12 +0000 Date: Tue, 14 Apr 2015 14:48:12 +0000 (UTC) From: "John Vines (JIRA)" To: notifications@accumulo.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (ACCUMULO-3349) LruBlockCache is miscalculating sizes MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/ACCUMULO-3349?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14494183#comment-14494183 ] John Vines commented on ACCUMULO-3349: -------------------------------------- I didn't, I'm not really sure how to test this since it is a concurrency bug. And I was only able to see this in LONG running systems (at least 6 weeks) so it's not something rapidly testable. And TBH, I'm not 100% certain that this is the fix. However, it was tracked down to this via code audit and even if this code isn't the culprit, the old logic was not correct, so at bare minimum that is fixed. > LruBlockCache is miscalculating sizes > ------------------------------------- > > Key: ACCUMULO-3349 > URL: https://issues.apache.org/jira/browse/ACCUMULO-3349 > Project: Accumulo > Issue Type: Bug > Affects Versions: 1.6.0 > Reporter: John Vines > Assignee: John Vines > Fix For: 1.7.0 > > Time Spent: 10m > Remaining Estimate: 0h > > On a system, we have index configured to 512MB and data block caches set to 1GB. I'm seeing the total size seemingly horrendously miscalculated, resulting in a LOT of eviction calls. > {code}2014-11-17 22:07:54,814 [cache.LruBlockCache] DEBUG: Block cache LRU eviction started. Attempting to free 58954734976 bytes > 2014-11-17 22:07:54,817 [cache.LruBlockCache] DEBUG: Block cache LRU eviction completed. Freed 102696 bytes. Priority Sizes: Single=186.14093MB (195182912), Multi=384.03125MB (402685944),Memory=0.0MB (0) > 2014-11-17 22:07:54,820 [cache.LruBlockCache] DEBUG: Block cache LRU eviction started. Attempting to free 58954734976 bytes > 2014-11-17 22:07:54,823 [cache.LruBlockCache] DEBUG: Block cache LRU eviction completed. Freed 102680 bytes. Priority Sizes: Single=186.14093MB (195182920), Multi=384.03125MB (402685936),Memory=0.0MB (0) > 2014-11-17 22:07:54,829 [cache.LruBlockCache] DEBUG: Block cache LRU eviction started. Attempting to free 58954734984 bytes > 2014-11-17 22:07:54,832 [cache.LruBlockCache] DEBUG: Block cache LRU eviction completed. Freed 205384 bytes. Priority Sizes: Single=186.043MB (195080224), Multi=384.12918MB (402788640),Memory=0.0MB (0) > {code} > {code}2014-11-17 22:06:39,439 [cache.LruBlockCache] DEBUG: Cache Stats: Sizes: Total=56991.715MB (59760144624), Free=-55967.715MB (-58686402800), Max=1024.0MB > (1073741824), Counts: Blocks=581902, Access=261753272, Hit=202812039, Miss=58941233, Evictions=45047347, Evicted=53777417, Ratios: Hit Ratio=77.4821341 > 0377502%, Miss Ratio=22.51785844564438%, Evicted/Run=1.1937975883483887, Duplicate Reads=4581914 > 2014-11-17 22:07:39,419 [cache.LruBlockCache] DEBUG: Cache Stats: Sizes: Total=835.4461MB (876028744), Free=-323.4461MB (-339157832), Max=512.0MB (536870912), Counts: Blocks=19881, Access=171866619, Hit=171174512, Miss=692107, Evictions=300619, Evicted=595529, Ratios: Hit Ratio=99.5972990989685%, Miss Ratio=0.402700062841177%, Evicted/Run=1.9810091257095337, Duplicate Reads=76697 > 2014-11-17 22:07:39,439 [cache.LruBlockCache] DEBUG: Cache Stats: Sizes: Total=56991.418MB (59759831656), Free=-55967.418MB (-58686089832), Max=1024.0MB (1073741824), Counts: Blocks=581901, Access=261755962, Hit=202813386, Miss=58942576, Evictions=45048521, Evicted=53778759, Ratios: Hit Ratio=77.48185992240906%, Miss Ratio=22.518140077590942%, Evicted/Run=1.1937963962554932, Duplicate Reads=4581916 > 2014-11-17 22:08:39,419 [cache.LruBlockCache] DEBUG: Cache Stats: Sizes: Total=835.1727MB (875742096), Free=-323.17273MB (-338871184), Max=512.0MB (536870912), Counts: Blocks=19762, Access=171879228, Hit=171187080, Miss=692148, Evictions=300642, Evicted=595689, Ratios: Hit Ratio=99.5972990989685%, Miss Ratio=0.40269438177347183%, Evicted/Run=1.9813898801803589, Duplicate Reads=76697 > {code} > Currently do not have log history to pinpoint where/where/if there was a slow incline to get to 55GB or if it was spontaneous at some discernible point. -- This message was sent by Atlassian JIRA (v6.3.4#6332)