From: Andreas Dilger Subject: Re: Problem with online resizing with metadata checksum patch... Date: Mon, 17 Sep 2012 18:42:50 -0400 Message-ID: References: Mime-Version: 1.0 (1.0) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8BIT Cc: "Darrick J. Wong" , "linux-ext4@vger.kernel.org" To: Theodore Ts'o Return-path: Received: from mail-qa0-f53.google.com ([209.85.216.53]:60362 "EHLO mail-qa0-f53.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755374Ab2IQWnE convert rfc822-to-8bit (ORCPT ); Mon, 17 Sep 2012 18:43:04 -0400 Received: by qaas11 with SMTP id s11so2041813qaa.19 for ; Mon, 17 Sep 2012 15:43:03 -0700 (PDT) In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On 2012-09-17, at 13:06, Theodore Ts'o wrote: > I don't know if you have any time to look at this, but if you do, that > would be great. > > I was testing online resizing with the metadata_csum feature, and found > that we're seeing a certain small amount of corrupted group descriptor > checksums. This doesn't occur with the standard original > uninit_bg/csum_gdt feature. > > This test below was done with the latest e2fsprogs next branch, as well > as the ext4 dev branch. I've also done a test with the ext4 origin > branch, which is ext4 with all of the patches that have gone upstream to > Linus's branch. The result was slightly different, but I saw the same > pattern of block group descriptor checksum failures and inode bitmap > corruption. > > If you could take a look at it, I'd really appreciate it!!! > > - Ted > > # mke2fs -t ext4 -O metadata_csum /dev/vdc 8M > # mount /dev/vdc /vdc > # resize2fs /dev/vdc 1024M > # umount /dev/vdc > # e2fsck -fy /dev/vdc > > e2fsck 1.43-WIP (31-Aug-2012) > One or more block group descriptor checksums are invalid. Fix? yes > > Group descriptor 31 checksum is 0xadcb, should be 0x9352. FIXED. > Group descriptor 63 checksum is 0xefc4, should be 0xf0db. FIXED. > Group descriptor 95 checksum is 0xa02c, should be 0x1d6b. FIXED. > Group descriptor 127 checksum is 0x4e35, should be 0xdf4a. FIXED. Looks like these are all multiples of 32 - 1? Cheers, Andreas > Pass 1: Checking inodes, blocks, and sizes > Pass 2: Checking directory structure > Pass 3: Checking directory connectivity > Pass 4: Checking reference counts > Pass 5: Checking group summary information > Block bitmap differences: -262144 > Fix? yes > > Inode bitmap differences: -(63510--63512) -63525 -(63542--63544) -63558 -(63574--63576) -(63590--63597) -63599 -63614 -63633 -(63681--63682) -63684 -(63686--63690) -63695 -63726 -(63731--63733) -63738 -63740 -63743 -63937 -(63941--63943) -63945 -63948 -63952 -64001 -(64022--64024) -64033 -64037 -(64054--64056) -64070 -64074 -(64086--64088) -64112 -64126 -(64145--64146) -64238 -(64241--64242) -64244 -(64246--64252) -64254 -64256 -64514 -(64534--64536) -64546 -64549 -(64566--64568) -64582 -64587 -(64598--64600) -64624 -64638 -(64657--64658) -64750 -(64753--64755) -(64758--64759) -(64761--64762) -(64764--64768) -(65025--65026) -(65046--65048) -(65057--65058) -65061 -(65078--65080) -65094 -(65098--65099) -(65110--65112) -65136 -65150 -(65169--65170) -65262 -(65267--65268) -65271 -65273 -652 75 -(65278--65280) -(129046--129048) -129061 -(129078--129080) -129094 -(129110--129112) -(129126--129133) -129135 -129150 -129169 -(129217--129218) -129220 -(129222--129226) -129231 -129262 - > (129267--129269) -129274 -129276 -129279 -129473 -(129477--129479) -129481 -129484 -129488 -129537 -(129558--129560) -129569 -129573 -(129590--129592) -129606 -129610 -(129622--129624) -129648 -129662 -(129681--129682) -129774 -(129777--129778) -129780 -(129782--129788) -129790 -129792 -130050 -(130070--130072) -130082 -130085 -(130102--130104) -130118 -130123 -(130134--130136) -130160 -130174 -(130193--130194) -130286 -(130289--130291) -(130294--130295) -(130297--130298) -(130300--130304) -(130561--130562) -(130582--130584) -(130593--130594) -130597 -(130614--130616) -130630 -(130634--130635) -(130646--130648) -130672 -130686 -(130705--130706) -130798 -(130803--130804) -130807 -130809 -130811 -(130814--130816) -(194582--194584) -194597 -(194614--194616) -194630 -(194646--194648) -(19466 2--194669) -194671 -194686 -194705 -(194753--194754) -194756 -(194758--194762) -194767 -194798 -(194803--194805) -194810 -194812 -194815 -195009 -(195013--195015) -195017 -195020 -195024 -195 > 073 -(195094--195096) -195105 -195109 -(195126--195128) -195142 -195146 -(195158--195160) -195184 -195198 -(195217--195218) -195310 -(195313--195314) -195316 -(195318--195324) -195326 -195328 -195586 -(195606--195608) -195618 -195621 -(195638--195640) -195654 -195659 -(195670--195672) -195696 -195710 -(195729--195730) -195822 -(195825--195827) -(195830--195831) -(195833--195834) -(195836--195840) -(196097--196098) -(196118--196120) -(196129--196130) -196133 -(196150--196152) -196166 -(196170--196171) -(196182--196184) -196208 -196222 -(196241--196242) -196334 -(196339--196340) -196343 -196345 -196347 -(196350--196352) -(260118--260120) -260133 -(260150--260152) -260166 -(260182--260184) -(260198--260205) -260207 -260222 -260241 -(260289--260290) -260292 -(260294--260298) -260303 -260334 -(260339--260341) -260346 -260348 -260351 -260545 -(260549--260551) -260553 -260556 -260560 -260609 -(260630--260632) -260641 -260645 -(260662--260664) -260678 -260682 -(260694--260696) -2607 > 20 -260734 -(260753--260754) -260846 -(260849--260850) -260852 -(260854--260860) -260862 -260864 -261122 -(261142--261144) -261154 -261157 -(261174--261176) -261190 -261195 -(261206--261208) -261232 -261246 -(261265--261266) -261358 -(261361--261363) -(261366--261367) -(261369--261370) -(261372--261376) -(261633--261634) -(261654--261656) -(261665--261666) -261669 -(261686--261688) -261702 -(261706--261707) -(261718--261720) -261744 -261758 -(261777--261778) -261870 -(261875--261876) -261879 -261881 -261883 -(261886--261888) > Fix? yes > > > /dev/vdc: ***** FILE SYSTEM WAS MODIFIED ***** > /dev/vdc: 11/262144 files (9.1% non-contiguous), 34426/1048576 blocks > candygram:~# > -- > To unsubscribe from this list: send the line "unsubscribe linux-ext4" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html