From mboxrd@z Thu Jan 1 00:00:00 1970 From: Edward Shishkin Subject: Re: Reiser4 extremely slow at deleting. Date: Wed, 27 Jul 2016 21:24:34 +0200 Message-ID: <57990A72.8070509@gmail.com> References: <0aa0ad49-a6a7-4d53-8a91-aa78821156c2@intersafeit.com> Mime-Version: 1.0 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=message-id:date:from:user-agent:mime-version:to:subject:references :in-reply-to:content-transfer-encoding; bh=Iuk5URkkopy6yqMTbmRteGs/KDMsJmSk3ycnPq4CIbg=; b=0hQgwX6aY2MGxgJURStRkDIxSfKJqZh4hqXtGmFqvvcqWANuzgPI1tbzbhvPpLhIwQ PYXE9w/rAGddZM+DfwUn1bO3m10BV2z2t88gkjOEEdonAr3KmQOCMuVrm+YNZm3tGM+P FPFx58/CqSN45Qgl6MGTF3QTVeLnhkFyUPOrB7QDy5ZH+OdUIlqwHFxETU8c3Q/pdnm7 n2rG3LteRLrsw/K24BJH/Br2al58mHiSZMfvlrn/tqfpkAyN2kFuFjxuUXJRKOEqTFqQ WdF19eTAssne4o5o82rA20F3c6m2vcunhHTp9dMlxEtMPcv5uhvA4T8thf4UWNQVLDf6 GaNA== In-Reply-To: Sender: reiserfs-devel-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="utf-8"; format="flowed" To: Georgios Tsalikis , reiserfs-devel@vger.kernel.org By default compression is turned on, and the file system tries to be=20 "smart". However, default heuristics is not perfect. It works well for "/"=20 (setups created during system installation). Also it works fine for various development= sets (mixes of well-squeezable sources and not squeezable binaries), which i= s=20 usual environment for Gentoo people. =46or media-files the default heuristics works very bad (is it clear, w= hy=20 so?). =46or large files it is strongly recommended to dedicate a separate=20 partition without compression (mkfs option "create=3Dreg40"). Thanks, Edward. On 07/27/2016 08:39 PM, Georgios Tsalikis wrote: > Here is my disappointing data :p > > # measurefs.reiser4 -p /dev/sdc3 > measurefs.reiser4 1.1.0 > Format release: 4.0.1 > Copyright (C) 2001-2005 by Hans Reiser, licensing governed by=20 > reiser4progs/COPYING. > > Default profile: > create: "ccreg40" (id:0x4 type:0x0) [Regular file= =20 > plugin for creat(2)] > key: "key_large" (id:0x1 type:0xb) [Key plugin] > node: "node40" (id:0x0 type:0x2) [Node plugin] > compress: "lzo1" (id:0x0 type:0xc) [Compression plu= gin] > compressMode: "conv" (id:0x4 type:0xd) [Compression Mod= e=20 > plugin] > cluster: "64K" (id:0x0 type:0x10) [Cluster plu= gin] > hash: "r5_hash" (id:0x1 type:0x3) [Directory=20 > entry hash plugin] > fibration: "ext_1_fibre" (id:0x2 type:0x4) [Key fibratio= n=20 > plugin] > formatting: "smart" (id:0x2 type:0x5) [File body=20 > formatting plugin] > > (actually i formatted it with node41 but it doesn't appear so) > > # ls -l resstick.dd > -rw-r--r-- 1 root root 22007840768 =CE=99=CE=BF=CF=8D=CE=BB 26 15:03 = resstick.dd > # time rm resstick.dd > > real 21m43.005s > user 0m0.000s > sys 5m2.132s > > If this is not problematic then what is? It is a single deletion!! > > > On 27/07/2016 09:15 =CE=BC=CE=BC, Morgan Smith wrote: >> Was the file system created using the reg40 plugin or ccreg40 for on= the >> fly compression? >> >> It seems to me I had this issue when using ccreg40 for on the fly >> compression. I wrote a similar request for help some years ago and I >> believe the answer was to not use compression. IIRC it's not an issu= e >> with the compression itself rather the handling of allocated space f= or >> it and the shuffling that is done when releasing resources. I was un= able >> to quickly find my old email to the list to confirm :( >> >> On 07/27/2016 11:51 AM, Georgios Tsalikis wrote: >>> It is almost always like this. For example a directory with 2.5GB o= f >>> large files takes about 2 minutes. Right now i am deleting a large >>> directory with small files, I didn't keep the numbers before runnin= g, >>> but they are many directories with source code. Could it be >>> fragmentation? And how could I mitigate it? Thanks >>> >>> >>> --=20 >>> To unsubscribe from this list: send the line "unsubscribe >>> reiserfs-devel" in >>> the body of a message to majordomo@vger.kernel.org >>> More majordomo info at http://vger.kernel.org/majordomo-info.html >> --=20 >> To unsubscribe from this list: send the line "unsubscribe=20 >> reiserfs-devel" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> > > > --=20 > To unsubscribe from this list: send the line "unsubscribe=20 > reiserfs-devel" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe reiserfs-deve= l" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html