LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Mandriva (https://www.linuxquestions.org/questions/mandriva-30/)
-   -   cp changes md5sum (https://www.linuxquestions.org/questions/mandriva-30/cp-changes-md5sum-340195/)

kel_en 07-05-2005 08:44 AM

cp changes md5sum
 
I am having an intermittent problem with data corruption as indicated by the md5sum changing for a file after copying it to another directory or other operations. I first noticed this after upgrading my system with an new motherboard, cpu and ram. The motherboard is a Soyo K7VMP2, the cpu is an AthlonXP 2800+ (Barton) and the memory is two PC2700 256M modules. The problem surfaced when building the Unichrome graphics drivers using the X source tree. I used ftp to obtain the matching xorg-x11 source SRPM from the mirror and installed it. Unzipping some of the .bz2 source files failed because of data corruption. Repeating the ftp and other operations ultimately resulted in a complete set of source files and I was able to successfully compile X without error and generate the 3-D drivers. But, the md5sum problem remains. When I upgraded the kernel from 2.6.8.1-24 to 2.6.8.1-25 I had to rebuild the Unichrome drivers and the whole file corruption problem began again.

I have Gentoo on another partition with the 2.6.11 kernel and I get the same md5sum changes in that partition, occasionally, when using the cp command on the xorg-x11 SRPM while booted into Gentoo.

I see no evidence of data corruption when running email, browser, OpenOffice, xpdf etc.. The system is perfectly stable with my primary partition, Mandrake 10.1. I have changed the md5sum of the xorg-x11 SRPM by doing cp, cp -a and rpm2cpio both as user and root. I don't recall if rpm-i changed the md5sum on xorg-x11 SRPM, but some of the files installed in /usr/src/RPM/SOURCES/ were clearly corrupted.

I have run memtest86 for hours without error. I have run cpuburn without error. Although the problem seemed to go away when I changed the IDE cable, it has since returned and I never see any errors in messages, syslog or dmesg. I have smartd running. My cpu die temperature is 54 degrees, idling, and 57 degrees under full load.

Any suggestions would be welcome.
Thanks.


All times are GMT -5. The time now is 12:44 PM.