lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <46E554B2.8060202@bull.net>
Date:	Mon, 10 Sep 2007 16:29:06 +0200
From:	Valerie Clement <valerie.clement@...l.net>
To:	Alex Tomas <alex@...sterfs.com>,
	Andreas Dilger <adilger@...sterfs.com>
Cc:	ext4 development <linux-ext4@...r.kernel.org>
Subject: file fragmentation with mballoc+delalloc options

There is another thing that looks strange to me.
Files created on an ext4 filesystem mounted with the mballoc and 
delalloc options seem to be more fragmented than those created on a 
filesystem mounted with only the delalloc option or mounted with only 
the mballoc option. Removing large files in the first case takes more 
time than in the 2 last cases.

Here are the small tests I did then (2.6.23-rc4 kernel + latest ext4 git 
patches):
# mkfs.ext3 -I 256 /dev/sdc
# mount -t ext4dev -o mballoc,delalloc,data=writeback /dev/sdc /mnt/test
# dd if=/dev/zero of=/mnt/test/foo bs=1M count=128
# sync
# filefrag /mnt/test/foo
   File is stored in extents format
   /mnt/test/foo: 62 extents found

# mkfs.ext3 -I 256 /dev/sdc
# mount -t ext4dev -o delalloc,data=writeback /dev/sdc /mnt/test
# dd if=/dev/zero of=/mnt/test/foo bs=1M count=128
# sync
# filefrag /mnt/test/foo
   File is stored in extents format
   /mnt/test/foo: 2 extents found

# mkfs.ext3 -I 256 /dev/sdc
# mount -t ext4dev -o mballoc,data=writeback /dev/sdc /mnt/test
# dd if=/dev/zero of=/mnt/test/foo bs=1M count=128
# sync
# filefrag /mnt/test/foo
   File is stored in extents format
   /mnt/test/foo: 2 extents found



Here is the output of "filefrag -v" when the filesystem is mounted with 
mballoc and delalloc options:

# filefrag -v /mnt/test/foo
Checking /mnt/test/foo
Filesystem type is: ef53
Filesystem cylinder groups is approximately 529
File is stored in extents format
Blocksize of file /mnt/test/foo is 4096
File size of /mnt/test/foo is 134217728 (32768 blocks)
First block: 37120
Last block: 72191
Discontinuity: Block 1024 is at 38400 (was 38143)
Discontinuity: Block 1036 is at 57344 (was 38411)
Discontinuity: Block 2048 is at 38412 (was 58355)
Discontinuity: Block 2072 is at 61440 (was 38435)
Discontinuity: Block 3072 is at 38436 (was 62439)
Discontinuity: Block 3108 is at 62440 (was 38471)
Discontinuity: Block 4096 is at 38472 (was 63427)
Discontinuity: Block 4144 is at 63428 (was 38519)
Discontinuity: Block 5120 is at 38520 (was 64403)
Discontinuity: Block 5180 is at 64404 (was 38579)
Discontinuity: Block 6144 is at 38580 (was 65367)
Discontinuity: Block 6216 is at 59392 (was 38651)
Discontinuity: Block 7168 is at 38652 (was 60343)
Discontinuity: Block 7252 is at 60344 (was 38735)
Discontinuity: Block 8192 is at 38736 (was 61283)
Discontinuity: Block 8288 is at 65368 (was 38831)
Discontinuity: Block 8448 is at 38144 (was 65527)
Discontinuity: Block 8704 is at 58624 (was 38399)
Discontinuity: Block 9216 is at 38912 (was 59135)
Discontinuity: Block 9324 is at 61284 (was 39019)
Discontinuity: Block 9472 is at 59136 (was 61431)
Discontinuity: Block 9728 is at 39424 (was 59391)
Discontinuity: Block 10240 is at 39020 (was 39935)
Discontinuity: Block 10360 is at 58369 (was 39139)
Discontinuity: Block 10496 is at 39936 (was 58504)
Discontinuity: Block 11264 is at 39140 (was 40703)
Discontinuity: Block 11396 is at 49152 (was 39271)
Discontinuity: Block 12288 is at 39272 (was 50043)
Discontinuity: Block 12432 is at 58505 (was 39415)
Discontinuity: Block 12544 is at 40704 (was 58616)
Discontinuity: Block 13468 is at 53248 (was 41627)
Discontinuity: Block 14336 is at 41628 (was 54115)
Discontinuity: Block 14504 is at 55296 (was 41795)
Discontinuity: Block 15360 is at 41796 (was 56151)
Discontinuity: Block 15540 is at 56152 (was 41975)
Discontinuity: Block 16384 is at 41984 (was 56995)
Discontinuity: Block 16576 is at 36928 (was 42175)
Discontinuity: Block 16640 is at 57088 (was 36991)
Discontinuity: Block 16896 is at 42496 (was 57343)
Discontinuity: Block 17408 is at 42176 (was 43007)
Discontinuity: Block 17612 is at 56996 (was 42379)
Discontinuity: Block 17664 is at 50176 (was 57047)
Discontinuity: Block 18432 is at 54272 (was 50943)
Discontinuity: Block 18648 is at 57048 (was 54487)
Discontinuity: Block 18688 is at 50944 (was 57087)
Discontinuity: Block 19456 is at 54488 (was 51711)
Discontinuity: Block 19684 is at 54116 (was 54715)
Discontinuity: Block 19712 is at 51712 (was 54143)
Discontinuity: Block 20480 is at 52736 (was 52479)
Discontinuity: Block 20720 is at 36912 (was 52975)
Discontinuity: Block 20736 is at 52480 (was 36927)
Discontinuity: Block 20992 is at 54784 (was 52735)
Discontinuity: Block 21504 is at 52976 (was 55295)
Discontinuity: Block 21756 is at 36908 (was 53227)
Discontinuity: Block 21760 is at 43008 (was 36911)
Discontinuity: Block 22792 is at 45056 (was 44039)
Discontinuity: Block 23808 is at 44040 (was 46071)
Discontinuity: Block 23828 is at 47104 (was 44059)
Discontinuity: Block 25856 is at 44544 (was 49131)
Discontinuity: Block 26368 is at 46080 (was 45055)
Discontinuity: Block 27392 is at 66816 (was 47103)
/mnt/test/foo: 62 extents found

What do you think about this?
   Valérie
-
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ