Red Hat Bugzilla – Bug 149608
CDs burnt from ISO images fail test (but seem ok)
Last modified: 2007-11-30 17:07:16 EST
From Bugzilla Helper: User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.1) Gecko/20040707 Description of problem: The CDs burnt from the ISO images downloaded from RedHat fail the 'media check' that should be done prior to installation. However, the CD seems fine. The MD5 checksum of the ISO image has been verified after download. The CD has been burnt on a W2K machine using Nero 5.5.10.56 with the options "finalize CD" OFF and the write method equal to "track-at-once". I have re-read the ISO image off the CD, which gives a file that nearly matches the original ISO image in size, but there is a 2KByte difference: 182151168 RHEL4-x86_64-ES-disc1.iso 182149120 Track.iso A hexdump and diff of these files reveals that the files differ at the end, the re-read iso image is missing 2KByte of 0s: $ tail RHEL4-x86_64-ES-disc1.iso.hex ad6a930 0000 0000 0000 0000 0000 0000 0000 0000 * ad6b000 2046 4f43 504d 5f53 4534 522e 4d50 313b ad6b010 2020 2020 2020 2020 2020 2020 2020 2020 * ad6b0d0 2020 6309 6d6f 7370 342d 5345 302d 322e ad6b0e0 3030 3035 3031 2e37 7273 2e63 7072 0a6d ad6b0f0 0000 0000 0000 0000 0000 0000 0000 0000 * adb6800 $ tail track.iso.hex ad6a930 0000 0000 0000 0000 0000 0000 0000 0000 * ad6b000 2046 4f43 504d 5f53 4534 522e 4d50 313b ad6b010 2020 2020 2020 2020 2020 2020 2020 2020 * ad6b0d0 2020 6309 6d6f 7370 342d 5345 302d 322e ad6b0e0 3030 3035 3031 2e37 7273 2e63 7072 0a6d ad6b0f0 0000 0000 0000 0000 0000 0000 0000 0000 * adb6000 So it seems that the media check is not lenient enough. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: See above Actual Results: See above Expected Results: See above Additional info:
Nero has been reported to actually misburn ISOs. I would recommend trying to find a port of cdrecord to do your burning with.
I have heard and tested that if you start linux with 'ide=nodma', the CDs actually pass. Which is nice. This seems to be a common problem btw. Should one not suggest to the user to start mediacheck with the dma disabled instead of just saying 'FAIL'? Here's an additional note I found, I did not test this btw: > I had the same problem until I made sure that I was burning "Disc At Once" > instead of "Track At Once" in Nero. > The default is "Track at Once," so be sure to change it to "Disc At Once" > every time you burn a new ISO. Now all 4 of my discs pass mediacheck, but I > haven't used them to install yet.
David, how are you reading the data off of the CD? You need to use "isosize" to find out how many bytes to read. If it really fails to read that number of bytes off of the CD, then nero is burning the CD incorrectly, and that is not a mediacheck bug. Unfortunately, the utility of "ide=nodma" almost varies according to what CD drive and IDE or SCSI controller you're using, and about as many drives will fail with ide=nodma as do without it.
Note that I have seem similar problems using cdrecord. This was due to a missing padding or an padsize that was too small. When using the -pad, the media check failed due to I/O errors at the end of the ISO image (fully reproducable with multiple CD's). When using a larger padsize, everything was ok. The failed CD's were fine for installing, b.t.w., but the mediacheck failed. One way to solve this problem is to add (enough) padding to the distributed ISO images after implanting the md5 checksum. This would save a lot of trouble for all people using bad CD burning programs and/or use too few padding space.