File corruption pattern
Neal Rhodes
neal at mnopltd.com
Wed May 18 19:04:19 UTC 2005
On Wed, May 18, 2005 at 10:14:07PM +1000, Michael D. Setzer II wrote:
> File corruption - Fedora Core 3
> shutdown -r -F now gives no errors.
>
> Image file created on a P4 system with no problem after
> earlier corruption on AMD 64 machine that was working before.
>
> 13,585,246,488 fc3ll.lzo
>
> File copied to AMD 64 machine using ncftp no error in copy,
> but then fails lzop -t <fc3ll.lzo test with checksum error.
>
> Then copied file to another system to compare original file to
> file from amd 64.
>
> System was working fine with 4 images, but now fails, on image
> creation,
> images that worked before, and even trying to blank drive with
> dd if=/dev/zero of=/0bits (Machine froze after about 8.3GB)
>
> Here is the comparision of the bad file versus good file.
> Note: all addresses end with 3 and all differences are either
> 08 hex or 80 hex?
>
> The system is less than 1 month old, and is an MSI 6741
> motherboard with a 250GB Maxtor Diamond Max 10 6B250R0 drive.
>
> Other than this error, which was working earlier with 4 pervious
> images, but now even those fail the lzop test.
>
> Ran the Gyum, and it went thru the install of the new updates with
> no problem.
>
> Any suggestions on what tests might help to find what is causing the
> problem. Does it look like motherboard, harddisk, hdparm, etc.
mmmmmmmmmmmmm, no, sorry.
We did chase problems with random file corruption on Redhat AS 3 using a
major brand name server with > 6GB of RAM a while back. Never resolved,
ended up sending back server and suing vendor.
WHAT WE LEARNED was that iozone was an excellent test. It has a -d
diagnostic option, which will compare what was written to what was read.
I'd suggest you run that with a test file of at least 3X your RAM or
16GB, whatever comes first. That's now a part of our standard server
test before putting anything on the box.
> Comparing files fc3ll.lzo and \FC3LL.LZO
> Address 64File P4 Diff
> 27B5C8B3: 62 E2 80
> 4BE758B3: 2C AC 80
> 4BE761A3: 64 6C 8
> 4BE768D3: 0E 8E 80
> 785F13C3: 05 0D 8
> 785F1943: E5 ED 8
> A1A788B3: 44 C4 80
> F3543943: F3 FB 8
> F35448D3: 6B EB 80
> FE52A1A3: B3 BB 8
> 171BB943: F6 FE 8
> 2ECBA8B3: 6D ED 80
> 4009F8B3: 4C CC 80
> 400A08D3: 17 97 80
> 8E3A71D3: 79 F9 80
> 8E3A73C3: E0 E8 8
> 8E3A78B3: 67 E7 80
> 92123823: 67 6F 8
> 921238B3: 6C EC 80
> BB3923C3: 23 2B 8
> BB392943: 11 19 8
> BCEA48D3: 3E BE 80
> BE538943: 27 2F 8
> BE5398D3: 6F EF 80
> E3BC71D3: 1A 9A 80
> E3BC73C3: 07 0F 8
> ECB0E1A3: 60 68 8
> 0CD6A8B3: 34 B4 80
> 0CD6A943: F7 FF 8
> 0E8898D3: 0B 8B 80
> 1C65F3C3: B0 B8 8
> 1C65F8B3: 71 F1 80
> 1C6C81A3: 13 1B 8
> 357BD8B3: 11 91 80
> 371BE1A3: 63 6B 8
> 4AB5E8B3: 50 D0 80
> 4AB5E943: F7 FF 8
> 78AB41A3: 01 09 8
> 78AB48D3: 45 C5 80
> 87EB63C3: 21 29 8
> 88B971A3: E2 EA 8
> 88B978D3: 4A CA 80
> A74861A3: C5 CD 8
> B11443C3: 70 78 8
> B1C9F1A3: 33 3B 8
> B1C9F8D3: 6F EF 80
> D5CC9943: 01 09 8
> DA7ED1A3: 21 29 8
> 03A658B3: 0A 8A 80
> 03A668D3: 61 E1 80
> 03AEF873: 6C EC 80
> 1A7B0143: 2E AE 80
> 1A7B03C3: 37 3F 8
> 1A7B0943: 24 2C 8
>
> Thanks.
>
> +----------------------------------------------------------+
> Michael D. Setzer II - Computer Science Instructor
> Guam Community College Computer Center
> mailto:mikes at kuentos.guam.net
> mailto:msetzerii at gmail.com
> http://www.guam.net/home/mikes
> Guam - Where America's Day Begins
> +----------------------------------------------------------+
>
> http://setiathome.berkeley.edu
> Number of Seti Units Returned: 16,494
> Processing time: 30 years, 322 days, 22 hours, 23 minutes
> (Total Hours: 270,550)
>
>
> --
> fedora-list mailing list
> fedora-list at redhat.com
> To unsubscribe: http://www.redhat.com/mailman/listinfo/fedora-list
--
============================================================================
Neal Rhodes MNOP Ltd (770) 972-5430
President 4737 Habersham Ridge fax: (770) 978-4741
Lilburn (atlanta) GA 30047
More information about the fedora-list
mailing list