Bug 169310

Summary:	corruption of larger files in software RAID 1
Product:	[Fedora] Fedora	Reporter:	Shafer Stockton <shafer>
Component:	kernel	Assignee:	Jeff Garzik <jgarzik>
Status:	CLOSED ERRATA	QA Contact:	Brian Brock <bbrock>
Severity:	medium	Docs Contact:
Priority:	medium
Version:	4	CC:	davej, jacob, peterm, wtogami
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2005-09-27 21:57:26 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Shafer Stockton 2005-09-26 21:10:07 UTC

Description of problem:

When copying four 600MB+ files (ISO's) from one section of the filesystem to
another at least one file is damaged.


Version-Release number of selected component (if applicable):

Seems to happen on all FC4 kernels and even the FC3 rescuecd kernel.


Steps to Reproduce:

1. I downloaded 4 ISO's via bittorrent (it took a while for the files to
checksum correctly, my first hint of problems)
2. Copy correctly checksummed files to another location on the FS (/tmp)
3. Check copies with sha1sum or md5sum (depending on the provided info for
correct checksums)

Additional info:

If I don't start the software RAID and interact with just half the mirror I can
copy files around without a problem.

I'm using software RAID 1 on two disk attached via the sata_via driver. The
motherboard is an ASUS A7V600-X with VIA VT8237 driving the Serial ATA. I tested
my memory for one full pass with memtest86 and things came out clean. I also ran
SeaTools full test on both my disks and they passed.

Comment 1 Dave Jones 2005-09-26 21:54:47 UTC

what is the newest kernel you've tried that exhibits this problem ?
There was a raid1 corruption problem fixed recently.

Comment 2 Shafer Stockton 2005-09-26 22:08:03 UTC

2.6.12-1.1456_FC4 which I'm fairly certain is the latest kernel (at least via
yum).  I tried it after I noticed the corruption with 2.6.12-1.1447_FC4.

Checking the testing directory just now I see a newer kernel. Is that the one
with the fix?

If it is I'll certainly try it. Guess I should have added CLOSED to my bugzilla
searches. Oops.

Comment 3 Dave Jones 2005-09-26 22:17:36 UTC

2.6.12-1.1456_FC4 has the fix I was thinking of, so this is a different problem.
Reassigning to SATA maintainer.

Comment 4 Shafer Stockton 2005-09-27 06:13:50 UTC

I had the fixed kernel installed and thought I had tested against it. Foolhardy
of me to think such things when I can't access the machine to double check
before replying.

The kernel is present and when used does indeed fix the corruption problem.

My apologies for crying wolf.

Comment 5 Dave Jones 2005-09-27 21:57:26 UTC

ok, thanks for double checking :-)