hi alexandr, this can be quite ok. if a cd is written, at the end of the cd there will be a "padding", so that if the laser reaches the end of the cd, the drive recognize it as the end (it is also called "lead-out") the length of that depends on the real end of the data, the lead out fills up the remaining space of the "sector" which can have different sizes. the lead-out aerea is normally not present in the source .iso file, only at the burned cd, as the software writes it "on the fly". hope thats right, it is the way someone explained that to me some years ago. greets, chris Am Freitag, 2. November 2001 12:03 schrieb Alexandr Malusek:
Hi,
I've noticed that the burning process sometimes results in a slightly shorter ISO image though the content of all burned files is the same as the content of the original files and no files are missing. Does anybody know why it happens? Can those CDs be considered OK?
More information: I burned 24 data CDs (cdrecord-1.9-54, SuSE 7.2, 2.4.4-64GB-SMP, IDE HP CD-Writer+ 9100, Kodak CD-R Ultima). Then I read these ISO images back via
dd if=/dev/cdrecorder of=xx_cd.iso
and compared them with the original images
cmp xx.iso xx_cd.iso
ISO images of 13 CDs matched the original (54%), 10 CDs had the ISO image shorter by exactly 2KB (otherwise OK), 1 ISO image was 4KB shorter (otherwise OK).
Then, for every CD, I compared (via cmp) each original file with the corresponding file on the CD . No difference was found.
There are no problems with reading those CDs so I keep them. But I'm a bit worried since they are used for archival purposes and some SW or HW problems may appear several years later.
I've seen the same problem on Solaris 2.5.1 using cdrecord and a Plextor SCSI writer. I've also tried CDs from other manufactures - there is some difference in the "success rate" but not extremely big.
-- Alexandr.Malusek@imv.liu.se
-- visit me at http://mamalala.de