From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 Description of problem: Summary ------- I suspect the kernel to corrupt the ext3 partition when intensive use of the disk is made (via cvs tagging). How reproducible ----------------- 3 disks corrupted in 2 months Chronology ---------- December 26: my RH7.2 system crashes --most of the 100GB disk can be recovered. I take the opportunity to upgrade to RH9 and a brand new Hitachi ATA/IDE 200GB disk. February 14: major slow down (the system is almost entirely un-responsive and X-windows doesn't refresh anymore), followed by a crash, while one of my users does a cvs tag ... After reboot, the disk shows substantial damage. One more reboot, and this time the disk can not be recovered at all. I buy another disk (WesternDigital WD2000), re-install RH9 and restore the data I had backed up (not all of it, alas, since I don't have a way to backup 200GB). February 24: Now, this new disk is starting to report errors as well: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error } hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=147238261, high=8, low=13020533, sector=142611536 end_request: I/O error, dev 03:03 (hda), sector 142611536 etc... Since this is the 3rd disk that ends up reporting the same type of error, each time while cvs is trying to update its repository, I am starting to suspect a software bug rather than a hardware one. Version-Release number of selected component (if applicable): kernel-2.4.20-28.9 How reproducible: Sometimes Steps to Reproduce: 1. cvs tag -rbuildxxxxxx product (most of the time works fine) 2. dma_intr problems appear in /var/log/messages 3. partition corrupted, sometimes boot won't even mount the partition anymore. Linux rescue won't mount it either. Additional info: FVIW, the project under cvs is 144MB (snapshot), while the cvs repository for that module is 809MB. Of course, since cvs is so clever, any tagging requires rewriting of all the files (just to add the tag at the beginning, so in our case that's 0.8GB every time).
Created attachment 98030 [details] output of dmesg
those errors do look very much like hardware failures unfortunatly. your ide controller isn't exactly uncommon either, so this would be more widespread if there was a bug there. I suggest you check cabling etc, and make sure you have a strong enough power supply.
Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/